NPN protective zener

[George Herold]

On Tuesday, June 25, 2019 at 5:04:26 PM UTC-4, John Larkin wrote:

On Tue, 25 Jun 2019 12:24:31 -0700 (PDT), George Herold
gherold@teachspin.com> wrote:

On Tuesday, June 25, 2019 at 3:16:46 PM UTC-4, John Larkin wrote:
On Tue, 25 Jun 2019 14:56:21 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 6/25/19 9:39 AM, tabbypurr@gmail.com wrote:
On Tuesday, 25 June 2019 03:28:50 UTC+1, John Larkin wrote:
On Mon, 24 Jun 2019 21:29:58 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:
On 6/21/19 8:09 PM, George Herold wrote:
On Friday, June 21, 2019 at 11:00:24 AM UTC-4, John Larkin wrote:

https://www.dropbox.com/s/ft0tsikhdi90rgq/BFT25.JPG?raw=1

The low end was limited by my measurement rig and may be even better.

Huh, can you say anything about the measurement rig.
(610 electrometer?)
I've measured current of about a pA with a ~10 pA bias current opamp
(difference measurement...

I think I was looking at c-b junction of a ~200V npn.
'HV' diode with low leakage.

The Keithley 610 is a pretty good instrument. Its bottom range is 10 fA
FS, which is more aspirational than functional, but the 100 fA range is
pretty solid, considering.

Some modern CMOS op amps will blow the 610 into the weeds.

Cheers

Phil Hobbs

My capacitor charging thing could get well into the aA range. I just
need a good SPDT switch.

Dracula's knife switch would at least give lots of clearance.


NT


Were his teeth bolted on?

Cheers

Phil Hobbs

(I suspect you meant Frankenstein.)


Only one side of one switch needs really low leakage.

https://www.dropbox.com/s/vj47r3uw9f3klm8/aA_Tester.JPG?raw=1

What's S1 for? (S2 resets cap and measures offset V.)
George H.


S1 isolates the integrating node from the opamp input current.

Thanks,
S2 ground-checks the opamp, to measure offset.

Pushing both discharges the cap.

How would we test this? The biggest resistor that I have is 1T. 1 volt
across that is 1,000,000 aA.

Test what? the cap? Discharge the cap with the 1T, and see it's
exponential... You want to measure the opamp bias current too?
The biggest R I have is 1 Gig. (1pA is 1mV which is easy on
a good DMM.)

George H.

--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

On Tue, 25 Jun 2019 12:19:22 -0700, John Larkin
<jjlarkin@highland_snip_technology.com> wrote:

On Tue, 25 Jun 2019 11:27:30 -0700, dplatt@coop.radagast.org (Dave
Platt) wrote:

In article <dfc2fa60-efca-4637-80ee-cf56dda879f5@googlegroups.com>,
tabbypurr@gmail.com> wrote:

My capacitor charging thing could get well into the aA range. I just
need a good SPDT switch.

Dracula's knife switch would at least give lots of clearance.

Wasn't it Dr. Frankenstein who had the big knife-switch?

And, if I recall correctly, that switch was always festooned with
cobwebs. This might create paths for excessive leakage, especially if
the weather is damp (which it usually was during the doctor's normal
work hours... something about thunderstorms...)



A small antenna and a high-z voltmeter does interesting things as a
cloud passes over. A lightning shot can really change the e-field.

When I was a kid, I had 40m and 20m dipole antennas strung across our
house, with coax fed down to my room. I'd always disconnect the
antennas when I wasn't using it. If there was any storm activity you
could hear the arcing across the PL259 connectors. At night it could
get really loud.

 

[whit3rd]

On Tuesday, June 25, 2019 at 2:04:26 PM UTC-7, John Larkin wrote:

How would we test this? The biggest resistor that I have is 1T. 1 volt
across that is 1,000,000 aA.

100 mv and resistor into an op amp with capacitor in feedback makes a ramp.
Ramp feeds second capacitor into test node.

Current through C2 is (C1 * 0.1)/(C2 *R)

A capacitance ratio of 10uf/100 pF would give you five orders of
magnitude in addition to the small-voltage-large-resistor
current. The 'burden' of the load/test capacitor complicates
the formula, but not by much.

Probably you want to integrate for a second or more, so a ramp
that goes 30V in three seconds would be... convenient.

 

> (1pA is 1mV which is easy on a good  DMM.)

The millivolt scale of a 10-Mohm meter reads picoamps directly.

Cheers

Phil Hobbs

 

[John Larkin]

On Thu, 27 Jun 2019 05:42:36 -0700 (PDT), pcdhobbs@gmail.com wrote:

 (1pA is 1mV which is easy on a good  DMM.)

The millivolt scale of a 10-Mohm meter reads picoamps directly.

Cheers

Phil Hobbs

Even cheap DVMs are pretty good at that.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Winfield Hill]

John Larkin wrote...

On 27 Jun 2019, pcdhobbs@gmail.com wrote:

 (1pA is 1mV which is easy on a good  DMM.)

The millivolt scale of a 10-Mohm meter reads picoamps directly.

Even cheap DVMs are pretty good at that.

I like the 0.001 nS conductance mode on
the Fluke 87 handheld multimeters.


--
Thanks,
- Win

 

On Tuesday, June 25, 2019 at 3:16:46 PM UTC-4, John Larkin wrote:

On Tue, 25 Jun 2019 14:56:21 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 6/25/19 9:39 AM, tabbypurr@gmail.com wrote:
On Tuesday, 25 June 2019 03:28:50 UTC+1, John Larkin wrote:
On Mon, 24 Jun 2019 21:29:58 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:
On 6/21/19 8:09 PM, George Herold wrote:
On Friday, June 21, 2019 at 11:00:24 AM UTC-4, John Larkin wrote:

https://www.dropbox.com/s/ft0tsikhdi90rgq/BFT25.JPG?raw=1

The low end was limited by my measurement rig and may be even better.

Huh, can you say anything about the measurement rig.
(610 electrometer?)
I've measured current of about a pA with a ~10 pA bias current opamp
(difference measurement...

I think I was looking at c-b junction of a ~200V npn.
'HV' diode with low leakage.

The Keithley 610 is a pretty good instrument. Its bottom range is 10 fA
FS, which is more aspirational than functional, but the 100 fA range is
pretty solid, considering.

Some modern CMOS op amps will blow the 610 into the weeds.

Cheers

Phil Hobbs

My capacitor charging thing could get well into the aA range. I just
need a good SPDT switch.

Dracula's knife switch would at least give lots of clearance.


NT


Were his teeth bolted on?

Cheers

Phil Hobbs

(I suspect you meant Frankenstein.)


Only one side of one switch needs really low leakage.

https://www.dropbox.com/s/vj47r3uw9f3klm8/aA_Tester.JPG?raw=1


That probably needs some home-made switch.

Why not dump the switches and just measure dV/dt?

Vdd
|
.---.
|DUT| Vdd
'---' |
| ||--' 2n7002 / BSS138
| ||<-.
+-----'|--+----->Vout
| |
--- |
--- 100pF (il)
| |
=== ===

You could bootstrap the FET's drain if those pF mattered.

Another approach could use a 10pF measurement cap, calibrate the
readings by adding a small calibrated cap in parallel, figure out
the capacitances, etc.

Or just run it with no integrator cap, then re-run adding in the 10pF.
Take the difference in slopes, and Attoamp's your uncle.

Cheers,
James Arthur

 

On Sunday, June 30, 2019 at 12:10:43 PM UTC-4, dagmarg...@yahoo.com wrote:

On Saturday, June 29, 2019 at 11:33:59 PM UTC-4, dagmarg...@yahoo.com wrote:
On Tuesday, June 25, 2019 at 3:16:46 PM UTC-4, John Larkin wrote:

Only one side of one switch needs really low leakage.

https://www.dropbox.com/s/vj47r3uw9f3klm8/aA_Tester.JPG?raw=1


That probably needs some home-made switch.

Why not dump the switches and just measure dV/dt?

Vdd
|
.---.
|DUT| Vdd
'---' |
| ||--' 2n7002 / BSS138
| ||<-.
+-----'|--+----->Vout
| |
--- |
--- 100pF (il)
| |
=== ===

You could bootstrap the FET's drain if those pF mattered.

Another approach could use a 10pF measurement cap, calibrate the
readings by adding a small calibrated cap in parallel, figure out
the capacitances, etc.

Or just run it with no integrator cap, then re-run adding in the 10pF.
Take the difference in slopes, and Attoamp's your uncle.

With the bootstrap, LTSpice figures this guy ramps at just a
tad (2%) under 1uV/S with 10aA input.

+12V +12V
| |
.---. '--| 2n5484
|DUT| .--|<--.
'---' | |
| ||--' BSS |
| ||<-. 123 |
+-----'|--+------+--->Vout
| |
--- |
--- 10pF \|2n3904
| |---+--22k---<+12
=== .<| |
| |
.-. V ~
1k | | --- red
'-' |
| |
-12V -12V

Without the bootstrap, Miller gobbles about half.

It doesn't have the precision of your switched approach,
but it's dirt-simple & permits continuous monitoring.

Um, 10e-18/10e-12 = microvolts per second, sorry. (fixed)

 

On Saturday, June 29, 2019 at 11:33:59 PM UTC-4, dagmarg...@yahoo.com wrote:

On Tuesday, June 25, 2019 at 3:16:46 PM UTC-4, John Larkin wrote:
On Tue, 25 Jun 2019 14:56:21 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 6/25/19 9:39 AM, tabbypurr@gmail.com wrote:
On Tuesday, 25 June 2019 03:28:50 UTC+1, John Larkin wrote:
On Mon, 24 Jun 2019 21:29:58 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:
On 6/21/19 8:09 PM, George Herold wrote:
On Friday, June 21, 2019 at 11:00:24 AM UTC-4, John Larkin wrote:

https://www.dropbox.com/s/ft0tsikhdi90rgq/BFT25.JPG?raw=1

The low end was limited by my measurement rig and may be even better.

Huh, can you say anything about the measurement rig.
(610 electrometer?)
I've measured current of about a pA with a ~10 pA bias current opamp
(difference measurement...

I think I was looking at c-b junction of a ~200V npn.
'HV' diode with low leakage.

The Keithley 610 is a pretty good instrument. Its bottom range is 10 fA
FS, which is more aspirational than functional, but the 100 fA range is
pretty solid, considering.

Some modern CMOS op amps will blow the 610 into the weeds.

Cheers

Phil Hobbs

My capacitor charging thing could get well into the aA range. I just
need a good SPDT switch.

Dracula's knife switch would at least give lots of clearance.


NT


Were his teeth bolted on?

Cheers

Phil Hobbs

(I suspect you meant Frankenstein.)


Only one side of one switch needs really low leakage.

https://www.dropbox.com/s/vj47r3uw9f3klm8/aA_Tester.JPG?raw=1


That probably needs some home-made switch.

Why not dump the switches and just measure dV/dt?

Vdd
|
.---.
|DUT| Vdd
'---' |
| ||--' 2n7002 / BSS138
| ||<-.
+-----'|--+----->Vout
| |
--- |
--- 100pF (il)
| |
=== ===

You could bootstrap the FET's drain if those pF mattered.

Another approach could use a 10pF measurement cap, calibrate the
readings by adding a small calibrated cap in parallel, figure out
the capacitances, etc.

Or just run it with no integrator cap, then re-run adding in the 10pF.
Take the difference in slopes, and Attoamp's your uncle.

With the bootstrap, LTSpice figures this guy ramps at just a
tad (2%) under 1mV/S with 10aA input.

+12V +12V
| |
.---. '--| 2n5484
|DUT| .--|<--.
'---' | |
| ||--' BSS |
| ||<-. 123 |
+-----'|--+------+--->Vout
| |
--- |
--- 10pF \|2n3904
| |---+--22k---<+12
=== .<| |
| |
.-. V ~>
1k | | --- red
'-' |
| |
-12V -12V

Without the bootstrap, Miller gobbles about half.

It doesn't have the precision of your switched approach,
but it's dirt-simple & permits continuous monitoring.

Cheers,
James Arthur

 

[John Larkin]

On Sun, 30 Jun 2019 09:10:39 -0700 (PDT), dagmargoodboat@yahoo.com
wrote:

On Saturday, June 29, 2019 at 11:33:59 PM UTC-4, dagmarg...@yahoo.com wrote:
On Tuesday, June 25, 2019 at 3:16:46 PM UTC-4, John Larkin wrote:
On Tue, 25 Jun 2019 14:56:21 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 6/25/19 9:39 AM, tabbypurr@gmail.com wrote:
On Tuesday, 25 June 2019 03:28:50 UTC+1, John Larkin wrote:
On Mon, 24 Jun 2019 21:29:58 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:
On 6/21/19 8:09 PM, George Herold wrote:
On Friday, June 21, 2019 at 11:00:24 AM UTC-4, John Larkin wrote:

https://www.dropbox.com/s/ft0tsikhdi90rgq/BFT25.JPG?raw=1

The low end was limited by my measurement rig and may be even better.

Huh, can you say anything about the measurement rig.
(610 electrometer?)
I've measured current of about a pA with a ~10 pA bias current opamp
(difference measurement...

I think I was looking at c-b junction of a ~200V npn.
'HV' diode with low leakage.

The Keithley 610 is a pretty good instrument. Its bottom range is 10 fA
FS, which is more aspirational than functional, but the 100 fA range is
pretty solid, considering.

Some modern CMOS op amps will blow the 610 into the weeds.

Cheers

Phil Hobbs

My capacitor charging thing could get well into the aA range. I just
need a good SPDT switch.

Dracula's knife switch would at least give lots of clearance.


NT


Were his teeth bolted on?

Cheers

Phil Hobbs

(I suspect you meant Frankenstein.)


Only one side of one switch needs really low leakage.

https://www.dropbox.com/s/vj47r3uw9f3klm8/aA_Tester.JPG?raw=1


That probably needs some home-made switch.

Why not dump the switches and just measure dV/dt?

Vdd
|
.---.
|DUT| Vdd
'---' |
| ||--' 2n7002 / BSS138
| ||<-.
+-----'|--+----->Vout
| |
--- |
--- 100pF (il)
| |
=== ===

You could bootstrap the FET's drain if those pF mattered.

Another approach could use a 10pF measurement cap, calibrate the
readings by adding a small calibrated cap in parallel, figure out
the capacitances, etc.

Or just run it with no integrator cap, then re-run adding in the 10pF.
Take the difference in slopes, and Attoamp's your uncle.

With the bootstrap, LTSpice figures this guy ramps at just a
tad (2%) under 1mV/S with 10aA input.

+12V +12V
| |
.---. '--| 2n5484
|DUT| .--|<--.
'---' | |
| ||--' BSS |
| ||<-. 123 |
+-----'|--+------+--->Vout
| |
--- |
--- 10pF \|2n3904
| |---+--22k---<+12
=== .<| |
| |
.-. V ~
1k | | --- red
'-' |
| |
-12V -12V

Without the bootstrap, Miller gobbles about half.

Nice circuit. The Vgs tempco will limit measuring really low currents.

The inherent time constant of an aA circuit makes these measurements
really tedious.

It doesn't have the precision of your switched approach,
but it's dirt-simple & permits continuous monitoring.

Cheers,
James Arthur

I don't know what the leakage of the fet might be; it could be aAs.

I played with a 2N7000 some time back:

https://www.dropbox.com/s/2mkh1au28f9utrs/2N7000.jpg?raw=1

As you can barely see, the gate is open. The fet is halfway turned on,
so the LED is half bright. It takes hours for the LED brightness to
change visibly. Gate leakage must be pretty low, very rough guess 5
aA, a few thousand electrons per second maybe.

One can tweak the LED current by touching the source or the drain with
a tiny insulated screwdriver, then touch the gate, to pump the current
up or down.

But given the abstract, arguably useless, challenge to get the lowest
possible current resolution, my switched cap thing is probably the
best way to do it.

A variation on the vibrating reed electrometer would be fun,
continuous readout with no lost charge.

I wonder how much charge just floats around in the air, ionization
somehow.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Winfield Hill]

John Larkin wrote...

I don't know what the leakage of the fet might be; it
could be aAs. I played with a 2N7000 some time back:

https://www.dropbox.com/s/2mkh1au28f9utrs/2N7000.jpg?raw=1

As you can barely see, the gate is open. The fet is
halfway turned on, so the LED is half bright. It takes
hours for the LED brightness to change visibly. ...

I made a set of open-gate measurements some time
ago, with a 5-digit meter measuring drain current.
It was amazingly stable, but with occasional jumps,
in the wrong direction for drain-gate currents.
Cosmic Rays?


--
Thanks,
- Win

 

[Winfield Hill]

dagmargoodboat@yahoo.com wrote...

With the bootstrap, LTSpice figures this guy ramps at
just a tad (2%) under 1uV/S with 10aA input.

+12V +12V
| |
.---. '--| 2n5484
|DUT| .--|<--.
'---' | |
| ||--' BSS |
| ||<-. 123 |
+-----'|--+------+--->Vout
| |
--- |
--- 10pF \|2n3904
| |---+--22k---<+12
=== .<| |
| |
.-. V ~
1k | | --- red
'-' |
| |
-12V -12V

Without the bootstrap, Miller gobbles about half.

It doesn't have the precision of your switched approach,
but it's dirt-simple & permits continuous monitoring.

Um, 10e-18/10e-12 = microvolts per second, sorry. (fixed)

A 2n5484 could be trouble, one we measured at 1mA showed
-Vgs approaching zero above 2mA, AoE III Figure 3.21.
Table 3.5 measurements show -Vgs = 0.73 volts for 2n5484
at 1mA, and suggest a 2n5485 or 5486 to help the MOSFET.
At any rate, bench measurements may be necessary.


--
Thanks,
- Win

 

[Jan Panteltje]

On a sunny day (30 Jun 2019 11:10:29 -0700) it happened Winfield Hill
<winfieldhill@yahoo.com> wrote in <qfatul0o51@drn.newsguy.com>:

John Larkin wrote...

I don't know what the leakage of the fet might be; it
could be aAs. I played with a 2N7000 some time back:

https://www.dropbox.com/s/2mkh1au28f9utrs/2N7000.jpg?raw=1

As you can barely see, the gate is open. The fet is
halfway turned on, so the LED is half bright. It takes
hours for the LED brightness to change visibly. ...

I made a set of open-gate measurements some time
ago, with a 5-digit meter measuring drain current.
It was amazingly stable, but with occasional jumps,
in the wrong direction for drain-gate currents.
Cosmic Rays?

Philips used this system in their TVs to hold the volume and picture settings,
they used a neon bulb as switch was lighted va a high value resistor
from some high negative or positive voltage to charge / discharge the capacitor,
Units were is a mall seeled plasict housing with the MOSFET and neon in it,
preventing cables and their noise and leakage to change the C voltage

Those things held their value for weeks, have not tried any longer.

Was posted here by me maybe > 10 years ago, is likely patented.

Something like this:
+
user buttons |
up -----------------|-----
| | | |
--- | | |
+100V 0 0 | |---- |
|--R 10M--|---(00)----||<--| |
-100V 0 0 | neon | |---| |
--- | | |------|--> to volume, contrast, brightness, colr etc controls,
| | === | | one such a sealed unit for each.
down | | [ ] |
| |_______| |
| | |
|_______________ |______|
|
///
Clever idea that neon, good insulator when not ignited.

It is old hat.

 

On Sunday, June 30, 2019 at 1:46:39 PM UTC-4, Winfield Hill wrote:

dagmargoodboat@yahoo.com wrote...

With the bootstrap, LTSpice figures this guy ramps at
just a tad (2%) under 1uV/S with 10aA input.

+12V +12V
| |
.---. '--| 2n5484
|DUT| .--|<--.
'---' | |
| ||--' BSS |
| ||<-. 123 |
+-----'|--+------+--->Vout
| |
--- |
--- 10pF \|2n3904
| |---+--22k---<+12
=== .<| |
| |
.-. V ~
1k | | --- red
'-' |
| |
-12V -12V

Without the bootstrap, Miller gobbles about half.

It doesn't have the precision of your switched approach,
but it's dirt-simple & permits continuous monitoring.

Um, 10e-18/10e-12 = microvolts per second, sorry. (fixed)

A 2n5484 could be trouble, one we measured at 1mA showed
-Vgs approaching zero above 2mA, AoE III Figure 3.21.
Table 3.5 measurements show -Vgs = 0.73 volts for 2n5484
at 1mA, and suggest a 2n5485 or 5486 to help the MOSFET.
At any rate, bench measurements may be necessary.

I have the advantage of specifying LTSpice(tm) brand JFETs;
accept no substitutes.

Yes, that's nicely spotted. Lowering the current sink value
would be another approach to ensure the MOSFET at least some
reasonable Vds.

I was trying to keep Vds low to minimize the potential idg
leakage. In retrospect, it might make more sense to center
Vg between Vd and Vs, geometry-weighted for the MOSFET, to
quasi-cancel any idg with an offsetting igs.

Cheers,
James Arthur

 

On Sunday, June 30, 2019 at 3:24:57 PM UTC-4, Jan Panteltje wrote:

On a sunny day (30 Jun 2019 11:10:29 -0700) it happened Winfield Hill
winfieldhill@yahoo.com> wrote in <qfatul0o51@drn.newsguy.com>:

John Larkin wrote...

I don't know what the leakage of the fet might be; it
could be aAs. I played with a 2N7000 some time back:

https://www.dropbox.com/s/2mkh1au28f9utrs/2N7000.jpg?raw=1

As you can barely see, the gate is open. The fet is
halfway turned on, so the LED is half bright. It takes
hours for the LED brightness to change visibly. ...

I made a set of open-gate measurements some time
ago, with a 5-digit meter measuring drain current.
It was amazingly stable, but with occasional jumps,
in the wrong direction for drain-gate currents.
Cosmic Rays?

Philips used this system in their TVs to hold the volume and picture settings,
they used a neon bulb as switch was lighted va a high value resistor
from some high negative or positive voltage to charge / discharge the capacitor,
Units were is a mall seeled plasict housing with the MOSFET and neon in it,
preventing cables and their noise and leakage to change the C voltage

Those things held their value for weeks, have not tried any longer.

Was posted here by me maybe > 10 years ago, is likely patented.

Something like this:
+
user buttons |
up -----------------|-----
| | | |
--- | | |
+100V 0 0 | |---- |
|--R 10M--|---(00)----||<--| |
-100V 0 0 | neon | |---| |
--- | | |------|--> to volume, contrast, brightness, colr etc controls,
| | === | | one such a sealed unit for each.
down | | [ ] |
| |_______| |
| | |
|_______________ |______|
|
///
Clever idea that neon, good insulator when not ignited.

It is old hat.

An updated version could use two Vgs LEDs in photovoltaic mode,
one up, one down.

Cheers,
James Arthur