Parasitic capacitance of SMD resistors and their generated noise...

[John Larkin]

On Mon, 7 Mar 2022 13:56:36 -0800 (PST), Phil Allison
<pallison49@gmail.com> wrote:

jla...@highlandsniptechnology.com wrote:
================================

Johnson noise must be the same regardless of the materials.

** A simple fact overwhealmed by persistent myth to the contrary.

Thick-film resistors can have excess noise if there is voltage across
them, but it\'s hard to measure.

** No it isn\'t.

https://tinyurl.com/y89rmtk3

First paragraph.

I managed to measure some once, but it
took very high value resistors and high voltages.

** Drivel.

Easy enough to characterise carbon film, carbon comp, metal glaze & MF types with values of say 100K and few DC volts.

Got some measurements?

>JL does not underdstand audio.

Well, my hearing is terrible, especially below 10 Hz.

--

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

 

[John Larkin]

On Mon, 7 Mar 2022 14:06:21 -0800 (PST), Phil Allison
<pallison49@gmail.com> wrote:

Phil Hobbs wrote:
==============

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.


** Really ????
That is massively ambiguous.

CC resistors can drop in value if they run hot and over time.
Mostly they slowly drift high, over a period of decades.

Crazy to crazy that \" non linearity\".


..... Phil

google resistor voltage coefficient

Carbons can be ballpark 100 PPM/volt.

--

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

 

[Phil Allison]

John Larkin wrote:
================

Johnson noise must be the same regardless of the materials.

** A simple fact overwhealmed by persistent myth to the contrary.

Thick-film resistors can have excess noise if there is voltage across
them, but it\'s hard to measure.

** No it isn\'t.

https://tinyurl.com/y89rmtk3

First paragraph.

** Yawnnnnnnn ....

How pathetic.

I managed to measure some once, but it
took very high value resistors and high voltages.

** Drivel.

Easy enough to characterise carbon film, carbon comp, metal glaze & MF types with values of say 100K and few DC volts.

Got some measurements?

** Published them years ago.

The excess noise as voltage was applied ( up to 20V) was marked, except for MF types

JL does not underdstand audio.

Well, my hearing is terrible,

** And all your other senses too.

especially below 10 Hz.

** Not audio.

Wot a tedious fuckhead.


...... Phil

 

[Phil Allison]

John Larkin bullshitted as usual wrote:
================================

Phil Hobbs wrote:
==============

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.


** Really ????

That is massively ambiguous.

CC resistors can drop in value if they run hot and over time.
Mostly they slowly drift high, over a period of decades.

Crazy to crazy that \" non linearity\".


google resistor voltage coefficient

** Google \" arrogant fuckwit \" - see yourself described.


> Carbons can be ballpark 100 PPM/volt.

** But are in fact not.


...... Phil

 

[John Larkin]

On Mon, 7 Mar 2022 15:07:49 -0800 (PST), Phil Allison
<pallison49@gmail.com> wrote:

John Larkin bullshitted as usual wrote:
================================

Phil Hobbs wrote:
==============

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.


** Really ????

That is massively ambiguous.

CC resistors can drop in value if they run hot and over time.
Mostly they slowly drift high, over a period of decades.

Crazy to crazy that \" non linearity\".


google resistor voltage coefficient

** Google \" arrogant fuckwit \" - see yourself described.


Carbons can be ballpark 100 PPM/volt.

** But are in fact not.


..... Phil

Comps can be 200.

In a tube amp with, say, a 50 volt p-p swing, that could be 10000 PPM
gain change, 1% distortion. That\'s almost enough to hear.

--

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

 

[John Larkin]

On Mon, 7 Mar 2022 14:53:48 -0800 (PST), Phil Allison
<pallison49@gmail.com> wrote:

John Larkin wrote:
================

Johnson noise must be the same regardless of the materials.

** A simple fact overwhealmed by persistent myth to the contrary.

Thick-film resistors can have excess noise if there is voltage across
them, but it\'s hard to measure.

** No it isn\'t.

https://tinyurl.com/y89rmtk3

First paragraph.

** Yawnnnnnnn ....

How pathetic.

I managed to measure some once, but it
took very high value resistors and high voltages.

** Drivel.

Easy enough to characterise carbon film, carbon comp, metal glaze & MF types with values of say 100K and few DC volts.

Got some measurements?

** Published them years ago.

The excess noise as voltage was applied ( up to 20V) was marked, except for MF types

JL does not underdstand audio.

Well, my hearing is terrible,

** And all your other senses too.

especially below 10 Hz.


** Not audio.

Wot a tedious fuckhead.

I\'m having fun. Are you?

--

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

 

[whit3rd]

On Monday, March 7, 2022 at 1:27:34 PM UTC-8, Phil Hobbs wrote:

whit3rd wrote:

Carbon resistors are semi-metals, so they get recombination noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

Semi-metals are semiconductors, but at or near their intrinsic temperatures.
A bunch of metal shavings in a tube, if shaken, has enough Shottky rectification
at the contact points to be an RF detector (coherer was the antique radio term).
Ceramics or metal oxides, with metal, are going to have minority carriers.
At each metal connection, you get carrier injection.

For metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so? I used to make conducting films of 100 angstroms or so. Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value. > So, carbon (carbon
film?) still is the solution for high-ohms items.

Carbon resistors, especially in presence of ozone, drift to higher values
because the carbon slowly turns to CO or CO2. Metals also grow oxide skins
(some more quickly than others). Substrate, metal, and glaze (or at least lacquer) plus
two solderable connections... a complete surface mount resistor packs a lot of
engineering into that little package.

 

[Phil Allison]

John Larkin bullshitted as usual wrote:
================================

Phil Hobbs wrote bullshit too:
==================

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.


** Really ????

That is massively ambiguous.

CC resistors can drop in value if they run hot and over time.
Mostly they slowly drift high, over a period of decades.

Crazy to crazy that \" non linearity\".


google resistor voltage coefficient

** Google \" arrogant fuckwit \" - see yourself described.


Carbons can be ballpark 100 PPM/volt.

** But are in fact not.


Comps can be 200.

** Bullshit

In a tube amp with, say, a 50 volt p-p swing, that could be 10000 PPM
gain change, 1% distortion.

** Made up, absolute crap !!!!
-----------------------------------------------
-----------------------------------------------

Just to refute JL\'s rampant insanity - I carried a test:
With a sine wave at 1KHz feeding a small toroidal supply tranny in reverse.
Max available = 250Vrms at 0.05% THD.

A pair of MF resistors ( 100k and 1k ) in series showed only residual THD across the 1k.
Change the 100k to a ( very old ) 0.5W CC and the THD reading was then 0.33%
Drop the input level to 50V rms and the reading was 0.07%.

250V rms = 705 V p-p.
50V rms = 141V p-p.

FYI

One cannot test the * lineartiy * using DC voltages.
Cos the damn things heat and drop value by several %.



...... Phil

 

[John S]

On 3/7/2022 3:27 PM, Phil Hobbs wrote:

whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination  noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For  metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so?  I used to make conducting films of 100 angstroms or so.  Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value.    > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs

Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.

 

[server]

On Mon, 7 Mar 2022 18:41:01 -0800 (PST), Phil Allison
<pallison49@gmail.com> wrote:

John Larkin bullshitted as usual wrote:
================================

Phil Hobbs wrote bullshit too:
==================

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.


** Really ????

That is massively ambiguous.

CC resistors can drop in value if they run hot and over time.
Mostly they slowly drift high, over a period of decades.

Crazy to crazy that \" non linearity\".


google resistor voltage coefficient

** Google \" arrogant fuckwit \" - see yourself described.


Carbons can be ballpark 100 PPM/volt.

** But are in fact not.


Comps can be 200.


** Bullshit

In a tube amp with, say, a 50 volt p-p swing, that could be 10000 PPM
gain change, 1% distortion.

** Made up, absolute crap !!!!
-----------------------------------------------
-----------------------------------------------

Just to refute JL\'s rampant insanity - I carried a test:
With a sine wave at 1KHz feeding a small toroidal supply tranny in reverse.
Max available = 250Vrms at 0.05% THD.

A pair of MF resistors ( 100k and 1k ) in series showed only residual THD across the 1k.
Change the 100k to a ( very old ) 0.5W CC and the THD reading was then 0.33%
Drop the input level to 50V rms and the reading was 0.07%.

250V rms = 705 V p-p.
50V rms = 141V p-p.

FYI

One cannot test the * lineartiy * using DC voltages.
Cos the damn things heat and drop value by several %.



..... Phil

I said \"comps\", meaning carbon composition. We were talking about
carbon resistors.





--

I yam what I yam - Popeye

 

[server]

On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination  noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For  metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so?  I used to make conducting films of 100 angstroms or so.  Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value.    > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.

In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.



--

I yam what I yam - Popeye

 

[John Walliker]

On Tuesday, 8 March 2022 at 10:51:26 UTC, jla...@highlandsniptechnology.com wrote:

On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sop...@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so? I used to make conducting films of 100 angstroms or so. Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value. > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.
In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.
--

I yam what I yam - Popeye

The long-term resistance shift doesn\'t behave that way at high voltages,
because the ones at the top of the string suffer more corona discharge which
can therefore erode them faster. I found this out the hard way.

John

 

[server]

On Tue, 8 Mar 2022 03:46:03 -0800 (PST), John Walliker
<jrwalliker@gmail.com> wrote:

On Tuesday, 8 March 2022 at 10:51:26 UTC, jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sop...@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so? I used to make conducting films of 100 angstroms or so. Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value. > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.
In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.
--

I yam what I yam - Popeye

The long-term resistance shift doesn\'t behave that way at high voltages,
because the ones at the top of the string suffer more corona discharge which
can therefore erode them faster. I found this out the hard way.

John

I guess it should be coated or potted or something.

Ww have a potential customer who wants a 1500 volt power supply
programmable and stable to 1 PPM. That could get interesting.

Some sort of shield could prevent corona. But 1500v isn\'t bad.





--

I yam what I yam - Popeye

 

[John S]

On 3/8/2022 4:51 AM, jlarkin@highlandsniptechnology.com wrote:

On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination  noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For  metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so?  I used to make conducting films of 100 angstroms or so.  Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value.    > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.

In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.

No, I don\'t think so. The resistors all have a negative voltage
coefficient and they add in a string.

 

[server]

On Tue, 8 Mar 2022 09:54:29 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/8/2022 4:51 AM, jlarkin@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination  noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For  metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so?  I used to make conducting films of 100 angstroms or so.  Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value.    > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.

In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.


No, I don\'t think so. The resistors all have a negative voltage
coefficient and they add in a string.

Given 10 identical resistors in series, each drops 1/10 of the total
voltage.



--

I yam what I yam - Popeye

 

[John S]

On 3/8/2022 10:04 AM, jlarkin@highlandsniptechnology.com wrote:

On Tue, 8 Mar 2022 09:54:29 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/8/2022 4:51 AM, jlarkin@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination  noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For  metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so?  I used to make conducting films of 100 angstroms or so.  Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty well.

AND keep it from oxidizing and changing value.    > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.

In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.


No, I don\'t think so. The resistors all have a negative voltage
coefficient and they add in a string.

Given 10 identical resistors in series, each drops 1/10 of the total
voltage.

Yes. But the resistance at rated voltage is much lower than it is at a
couple of volts.

 

[Phil Hobbs]

John S wrote:

On 3/8/2022 4:51 AM, jlarkin@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 21:18:05 -0600, John S <Sophi.2@invalid.org> wrote:

On 3/7/2022 3:27 PM, Phil Hobbs wrote:
whit3rd wrote:
On Monday, March 7, 2022 at 11:47:16 AM UTC-8, Phil Hobbs wrote:
John Larkin wrote:
On Mon, 7 Mar 2022 12:35:32 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

jla...@highlandsniptechnology.com wrote:
On Mon, 7 Mar 2022 12:03:17 -0500, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

...gizmo was exhibiting horrific 1/f noise on the spectrum
analyzer. It turned out to be the classic Tektronix 50-ohm 2-W
BNC feedthrough terminator I was using--it was made of cermet.
Switching to metal film dropped the 1/f noise by about two orders
of magnitude IIRC. Not subtle at all.

It\'s probably too late, but it would be interesting to measure the
1/f noise vs current. It might be square law, namely thermal.

Nah, definitely linear, and definitely 1/f not 1/f**2. A
constant-rate drift notionally has a spectrum that goes as 1/f**4,
and a random-walk drift (\'brown noise\') goes as 1/f**2.

Carbon resistors are semi-metals, so they get recombination  noise;

Not sure what you mean by that, exactly. To get recombination noise you
need minority carriers, no?

probably the \'cermet\' is a similar material.

Cermets aren\'t homogeneous--as the name implies, they\'re ceramic/metal
composites.

For  metal film, though, it\'s terribly difficult to get a thin and
long continuous metal path for high resistance,

Why so?  I used to make conducting films of 100 angstroms or so.  Atoms
are pretty small, and using sputtering as opposed to directional
evaporation will make the film follow even rough substrates pretty
well.

AND keep it from oxidizing and changing value.    > So, carbon (carbon
film?) still is the solution for high-ohms items.

It\'s certainly true that it\'s harder to make very high resistances out
of very low resistance materials, and there are lots of low-precision
applications where 1/f noise is not a serious issue--overvoltage
protection, for instance.

However, carbon resistors are seriously nonlinear at high voltages--the
resistance of old style Allen-Bradley carbon comps was allowed to drop
by a quarter at their upper voltage limit.

Cheers

Phil Hobbs


Yes! I learned that the hard way. Decided I was gonna make my own HV
divider from a string of 22meg CC resistors. I was astonished by the
change in resistance with voltage. So I made a new scale for the meter.

In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.


No, I don\'t think so. The resistors all have a negative voltage
coefficient and they add in a string.

But in a divider with 25% matched nonlinearity, you\'d have
1*0.75 / N*0.75 = 1 / N regardless.

Of course since the nonlinearity is caused by hot filaments beginning to
form inside the resistive element, one wouldn\'t expect them to be
particularly repeatable or time-invariant.

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

 

[Jan Panteltje]

On a sunny day (Tue, 08 Mar 2022 07:39:29 -0800) it happened
jlarkin@highlandsniptechnology.com wrote in
<5ute2hlgqblmtf9l7nf3kc8tasr8hooafq@4ax.com>:

I guess it should be coated or potted or something.

Ww have a potential customer who wants a 1500 volt power supply
programmable and stable to 1 PPM. That could get interesting.

Some sort of shield could prevent corona. But 1500v isn\'t bad.

Not just corona, any moisure in the air would screw things up, condensation on your PCB too.
Done a lot of HV CRT stuff, foucus was for example about 4 kV, done with a large carbon
wound resistor in the antique sets,,,
1 ppm is pushing it.
https://www.wagneronline.com.au/high-voltage-tv-focus-resistors/television-parts/service-repair-parts/1fcr150k-3714/679/pd/

 

[whit3rd]

On Tuesday, March 8, 2022 at 7:39:44 AM UTC-8, jla...@highlandsniptechnology.com wrote:

On Tue, 8 Mar 2022 03:46:03 -0800 (PST), John Walliker
jrwal...@gmail.com> wrote:

[about carbon resistors]

The long-term resistance shift doesn\'t behave that way at high voltages,
because the ones at the top of the string suffer more corona discharge which
can therefore erode them faster. I found this out the hard way.

I guess it should be coated or potted or something.

Ww have a potential customer who wants a 1500 volt power supply
programmable and stable to 1 PPM. That could get interesting.

Some sort of shield could prevent corona. But 1500v isn\'t bad.

It\'s about the same as focus voltages in old CRT displays; I\'ve replaced
a lot of focus resistors.

To prevent corona around a carbon resistor, you might consider getting the
resistor sealed in vacuum (glass tube package). At 45 kV auto ignition
wiring just had to go to 8mm diameter silicone insulation: the corona in
silicone is harder to start than corona in air, so... you build up the
insulator to fill the high-field volume entirely.

Victoreen and Ohmite make vacuum packages : <https://www.ebay.com/itm/123046764521>

 

[Phil Allison]

John Shithead wrote:

==================

In a divider made from a string of identical resistors, the
nonlinearity cancels. That\'s why the best dividers are all on the same
substrate.


No, I don\'t think so. The resistors all have a negative voltage
coefficient and they add in a string.

Given 10 identical resistors in series, each drops 1/10 of the total
voltage.

Yes.

** Game ove.,

But the resistance at rated voltage is much lower than it is at a
couple of volts.

** You awake ?

At any applied voltage, all the series Rs have the *same* voltage drop.
The 10:1 ratio is unchanging it they all have the same deviations.


....... Phil