an electronish puzzle

"Jim Thompson" <thegreatone@example.com> wrote in message
news:3sjp315oj88meq5gncgn1qkhdh5515c43u@4ax.com...
[referring to Brasfield]
And inexperienced. Anything with leads... leakage becomes totally
unpredictable.
Click to expand...
While I believe that to be irrelevent to the B-E reverse
breakdown puzzle, it goes to the diode leakage issue,
in which I have directly relevant experience.

Part of the reason I was willing to "stick my neck out"
is that I developed the core RC oscillator of a function
generator that was once sold commercially. (Not many,
as it turns out, but that instrument's performance was
not the reason!) If I had ready access to this resource
then, I would have been asking the same question.
As it happened, though, I had to do some research,
thinking, and experimentation to find a good way to
get accurate steering of pretty low currents. (The
low end frequency limit was 20 mHz or so.) My
solution, justified by the reasoning earlier posted,
some interesting relations between leakage bulk
resistivity, and breakdown voltage, as well as some
experiments with numerous parts, and finally use in
a stable RC slow oscillator, was to use the B-E
junction as suggested by my post in January.

You conjecture about experience, Jim. I can most
certainly concur with what you say about "leads",
(or, more to point, packaging), but the question
does not ultimately depend on that problem, one
which must be controlled at any rate when low
leakage is paramount. And that view comes from
real, hard experience, contrary to your suggestion.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
 
On Sat, 19 Mar 2005 14:43:28 -0800, "Larry Brasfield"
<donotspam_larry_brasfield@hotmail.com> wrote:

Hmmm, a slow day w.r.t. technical stuff.

Here is a question for anybody who claims (or would like
to pretend) expertise in semiconductor physics, and how
semiconductor device operation is explained or predicted.

Suppose you have a BJT of the usual sort, having
a low emitter-base reverse breakdown voltage.
Reverse bias the collector-base junction and put
a current meter in series with the collector. Now,
force a small current, in breakdown mode, thru
the emitter-base junction. What happens to the
current as indicated by the meter? Now, why?

Have fun!
Click to expand...


OK. Stuck in the lab, tweaking the tempco of a 50 MHz LC oscillator
(+125 ppm/k uncompensated), means I have time to kill. Lots of waiting
involved. So I tried it.

Fairchild 2N4400 fresh from the stock room (3 cents each). +10 on the
collector, base grounded, emitter to a power supply, all currents
measured.

At zero on the emitter supply, Ic is about 9 na; that's what you get
for 3 cents, I suppose. As Ve is pulled up, nothing happens until the
emitter zeners (or avalanches, whatever) and then Ic starts to
*increase*, contrary to, um, theory.

It's apparently linear up to Ie of 10 mA, all i did, at which time Ic
is up to 260 nA. The effective beta in this mode is then 2.6e-5.

So, why can't a poor engineer buy NTC surface mount ceramic capacitors
anywhere? That's the real mystery.


John
 
"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote
in message news:9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com...
On Sat, 19 Mar 2005 14:43:28 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:
....
Suppose you have a BJT of the usual sort, having
a low emitter-base reverse breakdown voltage.
Reverse bias the collector-base junction and put
a current meter in series with the collector. Now,
force a small current, in breakdown mode, thru
the emitter-base junction. What happens to the
current as indicated by the meter? Now, why?

OK. Stuck in the lab, tweaking the tempco of a 50 MHz LC oscillator
(+125 ppm/k uncompensated), means I have time to kill. Lots of waiting
involved. So I tried it.

Fairchild 2N4400 fresh from the stock room (3 cents each). +10 on the
collector, base grounded, emitter to a power supply, all currents
measured.

At zero on the emitter supply, Ic is about 9 na; that's what you get
for 3 cents, I suppose. As Ve is pulled up, nothing happens until the
emitter zeners (or avalanches, whatever) and then Ic starts to
*increase*, contrary to, um, theory.
Click to expand...
Neat.

I trust there was some real resistor limiting the
base-emitter breakdown current. I'm not sure
what would happen at "heat the part" levels.
And one of the fun parts of the question, (one
of those 2nd order effects I did not mention),
is that the collector-base leakage *should* go
up as you elevate that junction temperature,
a result sure to follow from your X current at
the 6 to 8 V defined by the base.

Also, the theory I offered did depend on tunnelling
at the emitter-base. I have never known whether
all transistors zener there, or whether some do and
others avalanche instead. That is why there is the
surely vague "a low emitter-base reverse breakdown
voltage". (Remember, this was an interview question.)
If the part you have is going into avalanche, there
should be a collector current increase because some
minority carriers manage to diffuse away from the
avalanche region, (at its edges). (The ones that do
not go the other way or turn into light.)

Do you have any RF transistors? (The kind with
the *really* low base-emitter breakdown voltage? ;-)
Those are much more likely to do zener mode.

It's apparently linear up to Ie of 10 mA, all i did, at which time Ic
is up to 260 nA. The effective beta in this mode is then 2.6e-5.
Click to expand...
It would be interesting to pulse the breakdown
current and see the short term response that is
not so influenced by junction temperature rises.

So, why can't a poor engineer buy NTC surface mount ceramic
capacitors anywhere? That's the real mystery.
Click to expand...
(Not trying to be cute, just a tip if needed ...)
Aren't they called COG or NPO now? Try:
http://www.mouser.com/?handler=data.listcategory&Ne=100&N=157

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
 
On Mon, 21 Mar 2005 11:55:26 -0800, "Larry Brasfield"
<donotspam_larry_brasfield@hotmail.com> wrote:

"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote
in message news:9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com...
On Sat, 19 Mar 2005 14:43:28 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:
...
Suppose you have a BJT of the usual sort, having
a low emitter-base reverse breakdown voltage.
Reverse bias the collector-base junction and put
a current meter in series with the collector. Now,
force a small current, in breakdown mode, thru
the emitter-base junction. What happens to the
current as indicated by the meter? Now, why?

OK. Stuck in the lab, tweaking the tempco of a 50 MHz LC oscillator
(+125 ppm/k uncompensated), means I have time to kill. Lots of waiting
involved. So I tried it.

Fairchild 2N4400 fresh from the stock room (3 cents each). +10 on the
collector, base grounded, emitter to a power supply, all currents
measured.

At zero on the emitter supply, Ic is about 9 na; that's what you get
for 3 cents, I suppose. As Ve is pulled up, nothing happens until the
emitter zeners (or avalanches, whatever) and then Ic starts to
*increase*, contrary to, um, theory.

Neat.

I trust there was some real resistor limiting the
base-emitter breakdown current.
Click to expand...
Oh, please.

I'm not sure
what would happen at "heat the part" levels.
Click to expand...
The numbers don't look like heating to me.

And one of the fun parts of the question, (one
of those 2nd order effects I did not mention),
is that the collector-base leakage *should* go
up as you elevate that junction temperature,
a result sure to follow from your X current at
the 6 to 8 V defined by the base.

Also, the theory I offered did depend on tunnelling
at the emitter-base. I have never known whether
all transistors zener there, or whether some do and
others avalanche instead. That is why there is the
surely vague "a low emitter-base reverse breakdown
voltage". (Remember, this was an interview question.)
Click to expand...
So, all sorts of guys didn't get hired because you expected them to
tell you something about transistors that's actually not true.

If the part you have is going into avalanche, there
should be a collector current increase because some
minority carriers manage to diffuse away from the
avalanche region, (at its edges). (The ones that do
not go the other way or turn into light.)

Do you have any RF transistors?
Click to expand...
Tons.

(The kind with
the *really* low base-emitter breakdown voltage? ;-)
Those are much more likely to do zener mode.

It's apparently linear up to Ie of 10 mA, all i did, at which time Ic
is up to 260 nA. The effective beta in this mode is then 2.6e-5.

It would be interesting to pulse the breakdown
current and see the short term response that is
not so influenced by junction temperature rises.
Click to expand...
Sounds like real work.


So, why can't a poor engineer buy NTC surface mount ceramic
capacitors anywhere? That's the real mystery.

(Not trying to be cute, just a tip if needed ...)
Aren't they called COG or NPO now? Try:
http://www.mouser.com/?handler=data.listcategory&Ne=100&N=157
Click to expand...

C0G/NP0 are close to zero TC. What is need is a net circuit
capacitance with a TC in the -250 PPM/K range, to compensate the
inductor, mostly. Most people want an order for hundreds of thousands
before they'll run off a batch.

Maybe these guys:

http://www.metcaps.com/Catalogue/Metuchen/Metuchen%2004.12.04/21.pdf



John
 
(Mostly a technical post)

"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com>
wrote in message news:5s9u31th9kin0kh5ouokdiu8d644jg39m2@4ax.com...
On Mon, 21 Mar 2005 11:55:26 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:
"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote
in message news:9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com...
On Sat, 19 Mar 2005 14:43:28 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:
...
Suppose you have a BJT of the usual sort, having
a low emitter-base reverse breakdown voltage.
Reverse bias the collector-base junction and put
a current meter in series with the collector. Now,
force a small current, in breakdown mode, thru
the emitter-base junction. What happens to the
current as indicated by the meter? Now, why?

OK. Stuck in the lab, tweaking the tempco of a 50 MHz LC oscillator
(+125 ppm/k uncompensated), means I have time to kill. Lots of waiting
involved. So I tried it.

Fairchild 2N4400 fresh from the stock room (3 cents each). +10 on the
collector, base grounded, emitter to a power supply, all currents
measured.

At zero on the emitter supply, Ic is about 9 na; that's what you get
for 3 cents, I suppose. As Ve is pulled up, nothing happens until the
emitter zeners (or avalanches, whatever) and then Ic starts to
*increase*, contrary to, um, theory.

Neat.

I trust there was some real resistor limiting the
base-emitter breakdown current.

Oh, please.
Click to expand...
Please, John. I expected so, but you know what
they say about spelling "assume". For all I knew,
you limited the current in some other way.

I'm not sure
what would happen at "heat the part" levels.

The numbers don't look like heating to me.
Click to expand...
With the temperature sensitivity of thermal generation
in Si, (about 3x per 10 oC), it would only take about
a 30 oC temperature rise to get the result you saw at
10 mA. For some 60 mW in a small part, I would
not be surprised if it had been a thermal effect. So I
wonder how linear was the current increase you saw?

And one of the fun parts of the question, (one
of those 2nd order effects I did not mention),
is that the collector-base leakage *should* go
up as you elevate that junction temperature,
a result sure to follow from your X current at
the 6 to 8 V defined by the base.

Also, the theory I offered did depend on tunnelling
at the emitter-base. I have never known whether
all transistors zener there, or whether some do and
others avalanche instead. That is why there is the
surely vague "a low emitter-base reverse breakdown
voltage". (Remember, this was an interview question.)
[non-electronics start]
So, all sorts of guys didn't get hired because you expected them to
tell you something about transistors that's actually not true.
Click to expand...
No, that is definitely not what happened. If that
was the way I operated, my managers should have
been fired for letting me do interviews at all.

My distinction between "posers" and "knowers" is
not the ego game some might imagine. Anybody
who does interviews and reads resumes knows
that they often are greatly exaggerated. One of
my tasks has long been to attempt to guage the
degree of such exaggeration.

There is real use for some ambiguities in interview
questions. The interviewee gets a chance to see
them and either ask a clarifying question, state
his/her assumption(s) when answering, or field a
question about it otherwise with the opportunity to
then show how they manage the problem of error,
however slight or serious it may be.

The purpose is to quickly evaluate a potential
hire, not to give her/him a pass/fail, or to "win".

[non-electronics end]

As for the "actually not true", that is not clear
for a couple reasons already suggested. (And
it is largely beside the point for the interview.
I have often been educated by interviewees.)

You wrote earlier:
It's apparently linear up to Ie of 10 mA, all i did,
at which time Ic is up to 260 nA.
If it was not exponential, as the temperature effect
would cause, then either your transistor is going
into avalanche or the mechanism at work is one
which has eluded everybody here (who played),
including myself, along with a number of bright
interviewees including one with a PhD obtained
after his work in novel semiconductor devices.
I will soon be asking him again about this puzzle.
For all I know, there could be another, parallel
effect in addition to all the tunneling.

Your numbers certainly suggest that only a small
fraction of the carriers injected are minority ones.
What you are seeing could well be one of the 2nd
order effects, beating out the one I mentioned. As
I have never attempted to quantify them, I am not
prepared to say that the 1st order one has been at
all disproved, even if your transistor was zenering.
As I mentioned, the reduction of minority carriers
in the base region is a pretty small effect, so one
should not be surprised if it is surpassed by one
or more of the other small effects.

If the part you have is going into avalanche, there
should be a collector current increase because some
minority carriers manage to diffuse away from the
avalanche region, (at its edges). (The ones that do
not go the other way or turn into light.)

Do you have any RF transistors?

Tons.
Click to expand...
That is a *lot* of transistors! So, can you get
interested in plugging one into your setup? (I
should assume not at this point, I guess.)

(The kind with
the *really* low base-emitter breakdown voltage? ;-)
Those are much more likely to do zener mode.

It's apparently linear up to Ie of 10 mA, all i did, at which time Ic
is up to 260 nA. The effective beta in this mode is then 2.6e-5.

It would be interesting to pulse the breakdown
current and see the short term response that is
not so influenced by junction temperature rises.

Sounds like real work.
Click to expand...
Sounded like an interesting experiment.

Another interesting experiment is to reverse the
transistor, (adjusting the 10V down, of course), and
breakdown the collector, while measuring emitter
current. That *will* be an avalanche scenario.

So, why can't a poor engineer buy NTC surface mount ceramic
capacitors anywhere? That's the real mystery.

(Not trying to be cute, just a tip if needed ...)
Aren't they called COG or NPO now? Try:
http://www.mouser.com/?handler=data.listcategory&Ne=100&N=157

C0G/NP0 are close to zero TC. What is need is a net circuit
capacitance with a TC in the -250 PPM/K range, to compensate the
inductor, mostly. Most people want an order for hundreds of thousands
before they'll run off a batch.
Click to expand...
(Blush.) I had completely forgotten about NTCs
and that characteristic. I have never used one,
and can but dimly remember classifying them as a
curiosity made for dialing-tone circuits or some
such niche. If you are a score-keeper, take a point
(or as many as you like). (Or keep it/them with
my "forced" concurrance. ;-)

Has anybody ever tried to build inductors with
very low TC? It seems that if the meter could
be a piece of metal for awhile, it should be
possible to make a stable coil. (Or is that TC
perfectly compensating one of the ferrites?)

Maybe these guys:
http://www.metcaps.com/Catalogue/Metuchen/Metuchen%2004.12.04/21.pdf
Good luck finding it.
Click to expand...

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
 
I read in sci.electronics.design that Larry Brasfield
<donotspam_larry_brasfield@hotmail.com> wrote (in
<wWI%d.25$Hp3.889@news.uswest.net>) about 'an electronish puzzle', on
Mon, 21 Mar 2005:

(Blush.) I had completely forgotten about NTCs and that
characteristic. I have never used one, and can but dimly remember
classifying them as a curiosity made for dialing-tone circuits or some
such niche.
Click to expand...
Back in the good old days, the *standard* low-value ceramic caps were of
N750 material. That, in due course, brought the matter of component TCs
to the attention of every newbie.
--
Regards, John Woodgate, OOO - Own Opinions Only.
There are two sides to every question, except
'What is a Moebius strip?'
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> a écrit dans le
message de news:5s9u31th9kin0kh5ouokdiu8d644jg39m2@4ax.com...
So, why can't a poor engineer buy NTC surface mount ceramic
capacitors anywhere? That's the real mystery.

(Not trying to be cute, just a tip if needed ...)
Aren't they called COG or NPO now? Try:
http://www.mouser.com/?handler=data.listcategory&Ne=100&N=157


C0G/NP0 are close to zero TC. What is need is a net circuit
capacitance with a TC in the -250 PPM/K range, to compensate the
inductor, mostly. Most people want an order for hundreds of thousands
before they'll run off a batch.
Click to expand...
Maybe you can buy a tin of violet paint?


--
Thanks,
Fred.
 
I read in sci.electronics.design that Fred Bartoli
<fred._canxxxel_this_bartoli@RemoveThatAlso_free.fr_AndThisToo> wrote
(in <423fe4ea$0$2048$636a15ce@news.free.fr>) about 'an electronish
puzzle', on Tue, 22 Mar 2005:
Maybe you can buy a tin of violet paint?

Violet is N750. OK, making 1/3 of the capacitance N750 and the rest
Click to expand...
NP0/COG would do. But yellow paint gets you N220. (;-)
--
Regards, John Woodgate, OOO - Own Opinions Only.
There are two sides to every question, except
'What is a Moebius strip?'
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
"John Woodgate" <jmw@jmwa.demon.contraspam.yuk> a écrit dans le message de
news:UhnOqPT9YAQCFwz5@jmwa.demon.co.uk...
I read in sci.electronics.design that Fred Bartoli
fred._canxxxel_this_bartoli@RemoveThatAlso_free.fr_AndThisToo> wrote
(in <423fe4ea$0$2048$636a15ce@news.free.fr>) about 'an electronish
puzzle', on Tue, 22 Mar 2005:
Maybe you can buy a tin of violet paint?

Violet is N750. OK, making 1/3 of the capacitance N750 and the rest
NP0/COG would do. But yellow paint gets you N220. (;-)
Click to expand...
So he has to get the missing 30ppm by carefully hand selecting NPO?

What a paint in the arse...


--
Thanks,
Fred.
 
I read in sci.electronics.design that Fred Bartoli
<fred._canxxxel_this_bartoli@RemoveThatAlso_free.fr_AndThisToo> wrote
(in <4240112b$0$2080$636a15ce@news.free.fr>) about 'an electronish
puzzle', on Tue, 22 Mar 2005:
"John Woodgate" <jmw@jmwa.demon.contraspam.yuk> a écrit dans le message de
news:UhnOqPT9YAQCFwz5@jmwa.demon.co.uk...
I read in sci.electronics.design that Fred Bartoli
fred._canxxxel_this_bartoli@RemoveThatAlso_free.fr_AndThisToo> wrote
(in <423fe4ea$0$2048$636a15ce@news.free.fr>) about 'an electronish
puzzle', on Tue, 22 Mar 2005:
Maybe you can buy a tin of violet paint?

Violet is N750. OK, making 1/3 of the capacitance N750 and the rest
NP0/COG would do. But yellow paint gets you N220. (;-)

So he has to get the missing 30ppm by carefully hand selecting NPO?

What a paint in the arse...


Actually, he still needs a sniff of N750 to *increase* the overall TC.
Click to expand...
But yes, it was a PITA correcting drift this way, especially as you had
to wait at least an hour to see the results of any change.
--
Regards, John Woodgate, OOO - Own Opinions Only.
There are two sides to every question, except
'What is a Moebius strip?'
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
On Mon, 21 Mar 2005 16:24:58 -0800, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Mon, 21 Mar 2005 15:39:06 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:


You wrote earlier:
It's apparently linear up to Ie of 10 mA, all i did,
at which time Ic is up to 260 nA.
If it was not exponential, as the temperature effect
would cause, then either your transistor is going
into avalanche or the mechanism at work is one
which has eluded everybody here (who played),
including myself, along with a number of bright
interviewees including one with a PhD obtained
after his work in novel semiconductor devices.
I will soon be asking him again about this puzzle.
For all I know, there could be another, parallel
effect in addition to all the tunneling.



I'm an engineer, so when theory collides with parts, I trust the
parts.


"One experiment is worth a thousand expert opinions."

- Werner Von Braun.
Click to expand...

---
Fred Bloggs wrote:

In the case of the OP292 thread, the motivation was to make a dead
certain diagnosis of the Nazi Clarence and his problem- and that is
what I did- also ordered some OP292's and played around with it some
more.
....


John Larkin wrote:

You actually ordered parts and built circuits to prove somebody from a
newsgroup to be wrong? That's pushing the top range of the
getalifeometer!

....


If you believe what you quoted:, "One experiment is worth a thousand
expert opinions.", then buying or building parts to prove a point in
the real world isn't something which should be ridiculed, it's
something which should be respected. After all, spending
discretionary income just to prove a point with no hope of
remuneration or, God forbid, gain, rocks. That is, it proves
committment to an ideal.

I seem to recall that you said that as a result of experimentation on
some of your stuff you're now shipping equipment with "oscilloscope
simulators" installed which, without, equipment you're selling won't
work properly. You also said that you didn't understand just why it
made the equipment work, so that seems to me to be a good example of
experiment taking the place of theory.

Anyway, it seems to me that you're unfairly chastising Fred for going
for empirical evidence of success since you're doing the same thing
and, eventually, it all has to work here, in the real world.

--
John Fields
 
In article <9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com>,
John Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:
[...]
So, why can't a poor engineer buy NTC surface mount ceramic capacitors
anywhere? That's the real mystery.
Click to expand...
SMT inductors often have a NTC. How about putting one in series with the
capacitor?


Most non-NPO capacitors have neg. voltage coef. So do SC cut crystals.

--
--
kensmith@rahul.net forging knowledge
 
On Mon, 21 Mar 2005 15:39:06 -0800, "Larry Brasfield"
<donotspam_larry_brasfield@hotmail.com> wrote:

(Mostly a technical post)

"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com
wrote in message news:5s9u31th9kin0kh5ouokdiu8d644jg39m2@4ax.com...
On Mon, 21 Mar 2005 11:55:26 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:
"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote
in message news:9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com...
On Sat, 19 Mar 2005 14:43:28 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:
...
Suppose you have a BJT of the usual sort, having
a low emitter-base reverse breakdown voltage.
Reverse bias the collector-base junction and put
a current meter in series with the collector. Now,
force a small current, in breakdown mode, thru
the emitter-base junction. What happens to the
current as indicated by the meter? Now, why?

OK. Stuck in the lab, tweaking the tempco of a 50 MHz LC oscillator
(+125 ppm/k uncompensated), means I have time to kill. Lots of waiting
involved. So I tried it.

Fairchild 2N4400 fresh from the stock room (3 cents each). +10 on the
collector, base grounded, emitter to a power supply, all currents
measured.

At zero on the emitter supply, Ic is about 9 na; that's what you get
for 3 cents, I suppose. As Ve is pulled up, nothing happens until the
emitter zeners (or avalanches, whatever) and then Ic starts to
*increase*, contrary to, um, theory.

Neat.

I trust there was some real resistor limiting the
base-emitter breakdown current.

Oh, please.

Please, John. I expected so, but you know what
they say about spelling "assume". For all I knew,
you limited the current in some other way.

I'm not sure
what would happen at "heat the part" levels.

The numbers don't look like heating to me.

With the temperature sensitivity of thermal generation
in Si, (about 3x per 10 oC), it would only take about
a 30 oC temperature rise to get the result you saw at
10 mA. For some 60 mW in a small part, I would
not be surprised if it had been a thermal effect. So I
wonder how linear was the current increase you saw?

And one of the fun parts of the question, (one
of those 2nd order effects I did not mention),
is that the collector-base leakage *should* go
up as you elevate that junction temperature,
a result sure to follow from your X current at
the 6 to 8 V defined by the base.

Also, the theory I offered did depend on tunnelling
at the emitter-base. I have never known whether
all transistors zener there, or whether some do and
others avalanche instead. That is why there is the
surely vague "a low emitter-base reverse breakdown
voltage". (Remember, this was an interview question.)
[non-electronics start]
So, all sorts of guys didn't get hired because you expected them to
tell you something about transistors that's actually not true.

No, that is definitely not what happened. If that
was the way I operated, my managers should have
been fired for letting me do interviews at all.

My distinction between "posers" and "knowers" is
not the ego game some might imagine. Anybody
who does interviews and reads resumes knows
that they often are greatly exaggerated. One of
my tasks has long been to attempt to guage the
degree of such exaggeration.

There is real use for some ambiguities in interview
questions. The interviewee gets a chance to see
them and either ask a clarifying question, state
his/her assumption(s) when answering, or field a
question about it otherwise with the opportunity to
then show how they manage the problem of error,
however slight or serious it may be.

The purpose is to quickly evaluate a potential
hire, not to give her/him a pass/fail, or to "win".

[non-electronics end]

As for the "actually not true", that is not clear
for a couple reasons already suggested. (And
it is largely beside the point for the interview.
I have often been educated by interviewees.)

You wrote earlier:
It's apparently linear up to Ie of 10 mA, all i did,
at which time Ic is up to 260 nA.
If it was not exponential, as the temperature effect
would cause, then either your transistor is going
into avalanche or the mechanism at work is one
which has eluded everybody here (who played),
including myself, along with a number of bright
interviewees including one with a PhD obtained
after his work in novel semiconductor devices.
I will soon be asking him again about this puzzle.
For all I know, there could be another, parallel
effect in addition to all the tunneling.

Your numbers certainly suggest that only a small
fraction of the carriers injected are minority ones.
What you are seeing could well be one of the 2nd
order effects, beating out the one I mentioned. As
I have never attempted to quantify them, I am not
prepared to say that the 1st order one has been at
all disproved, even if your transistor was zenering.
As I mentioned, the reduction of minority carriers
in the base region is a pretty small effect, so one
should not be surprised if it is surpassed by one
or more of the other small effects.

If the part you have is going into avalanche, there
should be a collector current increase because some
minority carriers manage to diffuse away from the
avalanche region, (at its edges). (The ones that do
not go the other way or turn into light.)

Do you have any RF transistors?

Tons.

That is a *lot* of transistors! So, can you get
interested in plugging one into your setup? (I
should assume not at this point, I guess.)

(The kind with
the *really* low base-emitter breakdown voltage? ;-)
Those are much more likely to do zener mode.

It's apparently linear up to Ie of 10 mA, all i did, at which time Ic
is up to 260 nA. The effective beta in this mode is then 2.6e-5.

It would be interesting to pulse the breakdown
current and see the short term response that is
not so influenced by junction temperature rises.

Sounds like real work.

Sounded like an interesting experiment.

Another interesting experiment is to reverse the
transistor, (adjusting the 10V down, of course), and
breakdown the collector, while measuring emitter
current. That *will* be an avalanche scenario.

So, why can't a poor engineer buy NTC surface mount ceramic
capacitors anywhere? That's the real mystery.

(Not trying to be cute, just a tip if needed ...)
Aren't they called COG or NPO now? Try:
http://www.mouser.com/?handler=data.listcategory&Ne=100&N=157

C0G/NP0 are close to zero TC. What is need is a net circuit
capacitance with a TC in the -250 PPM/K range, to compensate the
inductor, mostly. Most people want an order for hundreds of thousands
before they'll run off a batch.

(Blush.) I had completely forgotten about NTCs
and that characteristic. I have never used one,
and can but dimly remember classifying them as a
curiosity made for dialing-tone circuits or some
such niche. If you are a score-keeper, take a point
(or as many as you like). (Or keep it/them with
my "forced" concurrance. ;-)

Has anybody ever tried to build inductors with
very low TC? It seems that if the meter could
be a piece of metal for awhile, it should be
possible to make a stable coil. (Or is that TC
perfectly compensating one of the ferrites?)

Maybe these guys:
http://www.metcaps.com/Catalogue/Metuchen/Metuchen%2004.12.04/21.pdf
Good luck finding it.
Click to expand...
---
Instead of all the suggestions and cutesy bullshit, why don't you do
your own legwork and report back with what you find?

--
John Fields
 
On Tue, 22 Mar 2005 13:34:47 +0000, John Woodgate
<jmw@jmwa.demon.contraspam.yuk> wrote:

I read in sci.electronics.design that Fred Bartoli
fred._canxxxel_this_bartoli@RemoveThatAlso_free.fr_AndThisToo> wrote
(in <4240112b$0$2080$636a15ce@news.free.fr>) about 'an electronish
puzzle', on Tue, 22 Mar 2005:

"John Woodgate" <jmw@jmwa.demon.contraspam.yuk> a écrit dans le message de
news:UhnOqPT9YAQCFwz5@jmwa.demon.co.uk...
I read in sci.electronics.design that Fred Bartoli
fred._canxxxel_this_bartoli@RemoveThatAlso_free.fr_AndThisToo> wrote
(in <423fe4ea$0$2048$636a15ce@news.free.fr>) about 'an electronish
puzzle', on Tue, 22 Mar 2005:
Maybe you can buy a tin of violet paint?

Violet is N750. OK, making 1/3 of the capacitance N750 and the rest
NP0/COG would do. But yellow paint gets you N220. (;-)

So he has to get the missing 30ppm by carefully hand selecting NPO?

What a paint in the arse...


Actually, he still needs a sniff of N750 to *increase* the overall TC.
But yes, it was a PITA correcting drift this way, especially as you had
to wait at least an hour to see the results of any change.
Click to expand...

"Pain in the arse" is a good assessment. The current product uses a
varicap driven by an LM45 temp sensor, a couple of opamps, and a
trimpot. Even without soldering, adjusting the tc is a huge, tedious
nuisance. Now I want to make a really tiny version, so an NTC cap
would be ideal. There's a huge matrix of possibilities combining fixed
caps and NTCs; NTCs are theoretically available from N220 to N5600.
Disc NTC caps are still somewhat available, but it's hard to get
surface-mounts unless you want to buy a million, which makes
experimenting problematical.

If I could get the TC down from -125 to, say, +-40 maybe, it would
solve my problem.

John
 
John Fields wrote:
On Mon, 21 Mar 2005 15:39:06 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:


(Mostly a technical post)

"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com
wrote in message news:5s9u31th9kin0kh5ouokdiu8d644jg39m2@4ax.com...

On Mon, 21 Mar 2005 11:55:26 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:

"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote
in message news:9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com...

On Sat, 19 Mar 2005 14:43:28 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:

...

Suppose you have a BJT of the usual sort, having
a low emitter-base reverse breakdown voltage.
Reverse bias the collector-base junction and put
a current meter in series with the collector. Now,
force a small current, in breakdown mode, thru
the emitter-base junction. What happens to the
current as indicated by the meter? Now, why?

OK. Stuck in the lab, tweaking the tempco of a 50 MHz LC oscillator
(+125 ppm/k uncompensated), means I have time to kill. Lots of waiting
involved. So I tried it.

Fairchild 2N4400 fresh from the stock room (3 cents each). +10 on the
collector, base grounded, emitter to a power supply, all currents
measured.

At zero on the emitter supply, Ic is about 9 na; that's what you get
for 3 cents, I suppose. As Ve is pulled up, nothing happens until the
emitter zeners (or avalanches, whatever) and then Ic starts to
*increase*, contrary to, um, theory.

Neat.

I trust there was some real resistor limiting the
base-emitter breakdown current.

Oh, please.

Please, John. I expected so, but you know what
they say about spelling "assume". For all I knew,
you limited the current in some other way.


I'm not sure
what would happen at "heat the part" levels.

The numbers don't look like heating to me.

With the temperature sensitivity of thermal generation
in Si, (about 3x per 10 oC), it would only take about
a 30 oC temperature rise to get the result you saw at
10 mA. For some 60 mW in a small part, I would
not be surprised if it had been a thermal effect. So I
wonder how linear was the current increase you saw?


And one of the fun parts of the question, (one
of those 2nd order effects I did not mention),
is that the collector-base leakage *should* go
up as you elevate that junction temperature,
a result sure to follow from your X current at
the 6 to 8 V defined by the base.

Also, the theory I offered did depend on tunnelling
at the emitter-base. I have never known whether
all transistors zener there, or whether some do and
others avalanche instead. That is why there is the
surely vague "a low emitter-base reverse breakdown
voltage". (Remember, this was an interview question.)

[non-electronics start]

So, all sorts of guys didn't get hired because you expected them to
tell you something about transistors that's actually not true.

No, that is definitely not what happened. If that
was the way I operated, my managers should have
been fired for letting me do interviews at all.

My distinction between "posers" and "knowers" is
not the ego game some might imagine. Anybody
who does interviews and reads resumes knows
that they often are greatly exaggerated. One of
my tasks has long been to attempt to guage the
degree of such exaggeration.

There is real use for some ambiguities in interview
questions. The interviewee gets a chance to see
them and either ask a clarifying question, state
his/her assumption(s) when answering, or field a
question about it otherwise with the opportunity to
then show how they manage the problem of error,
however slight or serious it may be.

The purpose is to quickly evaluate a potential
hire, not to give her/him a pass/fail, or to "win".

[non-electronics end]

As for the "actually not true", that is not clear
for a couple reasons already suggested. (And
it is largely beside the point for the interview.
I have often been educated by interviewees.)

You wrote earlier:
It's apparently linear up to Ie of 10 mA, all i did,
at which time Ic is up to 260 nA.
If it was not exponential, as the temperature effect
would cause, then either your transistor is going
into avalanche or the mechanism at work is one
which has eluded everybody here (who played),
including myself, along with a number of bright
interviewees including one with a PhD obtained
after his work in novel semiconductor devices.
I will soon be asking him again about this puzzle.
For all I know, there could be another, parallel
effect in addition to all the tunneling.

Your numbers certainly suggest that only a small
fraction of the carriers injected are minority ones.
What you are seeing could well be one of the 2nd
order effects, beating out the one I mentioned. As
I have never attempted to quantify them, I am not
prepared to say that the 1st order one has been at
all disproved, even if your transistor was zenering.
As I mentioned, the reduction of minority carriers
in the base region is a pretty small effect, so one
should not be surprised if it is surpassed by one
or more of the other small effects.


If the part you have is going into avalanche, there
should be a collector current increase because some
minority carriers manage to diffuse away from the
avalanche region, (at its edges). (The ones that do
not go the other way or turn into light.)

Do you have any RF transistors?

Tons.

That is a *lot* of transistors! So, can you get
interested in plugging one into your setup? (I
should assume not at this point, I guess.)


(The kind with
the *really* low base-emitter breakdown voltage? ;-)
Those are much more likely to do zener mode.


It's apparently linear up to Ie of 10 mA, all i did, at which time Ic
is up to 260 nA. The effective beta in this mode is then 2.6e-5.

It would be interesting to pulse the breakdown
current and see the short term response that is
not so influenced by junction temperature rises.

Sounds like real work.

Sounded like an interesting experiment.

Another interesting experiment is to reverse the
transistor, (adjusting the 10V down, of course), and
breakdown the collector, while measuring emitter
current. That *will* be an avalanche scenario.


So, why can't a poor engineer buy NTC surface mount ceramic
capacitors anywhere? That's the real mystery.

(Not trying to be cute, just a tip if needed ...)
Aren't they called COG or NPO now? Try:
http://www.mouser.com/?handler=data.listcategory&Ne=100&N=157

C0G/NP0 are close to zero TC. What is need is a net circuit
capacitance with a TC in the -250 PPM/K range, to compensate the
inductor, mostly. Most people want an order for hundreds of thousands
before they'll run off a batch.

(Blush.) I had completely forgotten about NTCs
and that characteristic. I have never used one,
and can but dimly remember classifying them as a
curiosity made for dialing-tone circuits or some
such niche. If you are a score-keeper, take a point
(or as many as you like). (Or keep it/them with
my "forced" concurrance. ;-)

Has anybody ever tried to build inductors with
very low TC? It seems that if the meter could
be a piece of metal for awhile, it should be
possible to make a stable coil. (Or is that TC
perfectly compensating one of the ferrites?)


Maybe these guys:
http://www.metcaps.com/Catalogue/Metuchen/Metuchen%2004.12.04/21.pdf

Good luck finding it.


---
Instead of all the suggestions and cutesy bullshit, why don't you do
your own legwork and report back with what you find?
Click to expand...
Oh didn't ya' know? You're supposed to consider yourself *privileged* to
receive direction from Larry Brasfield. Seriously, the p.o.s. *has* to
be posting from an asylum somewhere.
 
"John Fields" <jfields@austininstruments.com> wrote in message
news:aoe0411sep2qt5mcomjjel04u5lgslrmbg@4ax.com...
[SNIP]
, why don't you do
your own legwork and report back with what you find?
Click to expand...
I plan to do just that, presuming you mean variations on the
experiment reported by John Larkin. AC b-e current is
sure to separate thermal effects from the other effects.

--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
 
On Tue, 22 Mar 2005 09:13:25 -0600, John Fields
<jfields@austininstruments.com> wrote:

On Mon, 21 Mar 2005 16:24:58 -0800, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Mon, 21 Mar 2005 15:39:06 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:


You wrote earlier:
It's apparently linear up to Ie of 10 mA, all i did,
at which time Ic is up to 260 nA.
If it was not exponential, as the temperature effect
would cause, then either your transistor is going
into avalanche or the mechanism at work is one
which has eluded everybody here (who played),
including myself, along with a number of bright
interviewees including one with a PhD obtained
after his work in novel semiconductor devices.
I will soon be asking him again about this puzzle.
For all I know, there could be another, parallel
effect in addition to all the tunneling.



I'm an engineer, so when theory collides with parts, I trust the
parts.


"One experiment is worth a thousand expert opinions."

- Werner Von Braun.


---
Fred Bloggs wrote:

In the case of the OP292 thread, the motivation was to make a dead
certain diagnosis of the Nazi Clarence and his problem- and that is
what I did- also ordered some OP292's and played around with it some
more.
...


John Larkin wrote:

You actually ordered parts and built circuits to prove somebody from a
newsgroup to be wrong? That's pushing the top range of the
getalifeometer!

...


If you believe what you quoted:, "One experiment is worth a thousand
expert opinions.", then buying or building parts to prove a point in
the real world isn't something which should be ridiculed, it's
something which should be respected. After all, spending
discretionary income just to prove a point with no hope of
remuneration or, God forbid, gain, rocks. That is, it proves
committment to an ideal.
Click to expand...
The "ideal", in his case, is to prove other people stupid in cases
where ordinary insults and obscenities aren't enough.

If Fred had a legitimate need, or just intellectual curiosity, about
this issue, then that has my respect; so why doesn't he share it with
us?

If he did it to prove that someone is an idiot, then he's running true
to form.

I seem to recall that you said that as a result of experimentation on
some of your stuff you're now shipping equipment with "oscilloscope
simulators" installed which, without, equipment you're selling won't
work properly.
Click to expand...
The "oscilloscope simulators" part doesn't sound familiar to me.

You also said that you didn't understand just why it
made the equipment work, so that seems to me to be a good example of
experiment taking the place of theory.
Click to expand...
Can't recall that. Got a reference?

Oh, do you mean the cap on the Xilinx pin? That's hardly an
oscilloscope simulator! It's zapping some ringing or something. You
might note that I shared that observation with the OP in the hopes of
helping him with a similar problem, and did *not* insult him or call
him an idiot.

Theory is only useful if it's predictive; even the scientists agree on
that. I just finished reading Dahl's book, "Flash of the Cathode
Rays", which is all about this. Surprisingly good, fun to read book,
given that it's about the history of subatomic particles. The parts
about N-rays and sub-e charges are especially cool.

Anyway, it seems to me that you're unfairly chastising Fred for going
for empirical evidence of success since you're doing the same thing
and, eventually, it all has to work here, in the real world.
Click to expand...
I chastise Fred for heaping extended, crude, obscene, and frequent
insults on the majority of the posters in this group, including
newbies who ask innocent (if often naive) questions. He's not here to
learn, or to help, or to have fun, but mostly just to show how smart
he is and how stupid everybody else is. I wonder what the real story
is of Fred's life; he's silent on that.

For pete's sake, read his posts. He's rude, obscene, and perfectly
mean-spirited. If that's the sort of person you admire, well...


John
 
On Tue, 22 Mar 2005 09:38:39 -0600, John Fields
<jfields@austininstruments.com> wrote:


Instead of all the suggestions and cutesy bullshit, why don't you do
your own legwork and report back with what you find?
Click to expand...
Because this is a discussion group, not a scientific journal. And
besides, somebody who knows a lot about this may have ideas that could
help me.

John
 
On Tue, 22 Mar 2005 15:16:50 +0000 (UTC), kensmith@green.rahul.net
(Ken Smith) wrote:

In article <9a4u319497u9jbn4ljfk5afjobp7oc3dq8@4ax.com>,
John Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:
[...]
So, why can't a poor engineer buy NTC surface mount ceramic capacitors
anywhere? That's the real mystery.

SMT inductors often have a NTC. How about putting one in series with the
capacitor?
Click to expand...

The inductor I'm using in this oscillator, a 1008 surface-mount, has
(apparently) a strong positive TC. All the aircore inductors I've used
had positive TCs; I guess they just expand with temperature, and
bigger diameter makes more L. This is only 150 nH, so a
temperature-compensating core is probably out of the question.


John
 
On Tue, 22 Mar 2005 08:34:35 -0800, John Larkin
<jjSNIPlarkin@highTHISlandPLEASEtechnology.XXX> wrote:

On Tue, 22 Mar 2005 09:13:25 -0600, John Fields
jfields@austininstruments.com> wrote:

On Mon, 21 Mar 2005 16:24:58 -0800, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Mon, 21 Mar 2005 15:39:06 -0800, "Larry Brasfield"
donotspam_larry_brasfield@hotmail.com> wrote:


You wrote earlier:
It's apparently linear up to Ie of 10 mA, all i did,
at which time Ic is up to 260 nA.
If it was not exponential, as the temperature effect
would cause, then either your transistor is going
into avalanche or the mechanism at work is one
which has eluded everybody here (who played),
including myself, along with a number of bright
interviewees including one with a PhD obtained
after his work in novel semiconductor devices.
I will soon be asking him again about this puzzle.
For all I know, there could be another, parallel
effect in addition to all the tunneling.



I'm an engineer, so when theory collides with parts, I trust the
parts.


"One experiment is worth a thousand expert opinions."

- Werner Von Braun.


---
Fred Bloggs wrote:

In the case of the OP292 thread, the motivation was to make a dead
certain diagnosis of the Nazi Clarence and his problem- and that is
what I did- also ordered some OP292's and played around with it some
more.
...


John Larkin wrote:

You actually ordered parts and built circuits to prove somebody from a
newsgroup to be wrong? That's pushing the top range of the
getalifeometer!

...


If you believe what you quoted:, "One experiment is worth a thousand
expert opinions.", then buying or building parts to prove a point in
the real world isn't something which should be ridiculed, it's
something which should be respected. After all, spending
discretionary income just to prove a point with no hope of
remuneration or, God forbid, gain, rocks. That is, it proves
committment to an ideal.


The "ideal", in his case, is to prove other people stupid in cases
where ordinary insults and obscenities aren't enough.

If Fred had a legitimate need, or just intellectual curiosity, about
this issue, then that has my respect; so why doesn't he share it with
us?

If he did it to prove that someone is an idiot, then he's running true
to form.

I seem to recall that you said that as a result of experimentation on
some of your stuff you're now shipping equipment with "oscilloscope
simulators" installed which, without, equipment you're selling won't
work properly.

The "oscilloscope simulators" part doesn't sound familiar to me.

You also said that you didn't understand just why it
made the equipment work, so that seems to me to be a good example of
experiment taking the place of theory.

Can't recall that. Got a reference?

Oh, do you mean the cap on the Xilinx pin? That's hardly an
oscilloscope simulator! It's zapping some ringing or something. You
might note that I shared that observation with the OP in the hopes of
helping him with a similar problem, and did *not* insult him or call
him an idiot.

Theory is only useful if it's predictive; even the scientists agree on
that. I just finished reading Dahl's book, "Flash of the Cathode
Rays", which is all about this. Surprisingly good, fun to read book,
given that it's about the history of subatomic particles. The parts
about N-rays and sub-e charges are especially cool.

Anyway, it seems to me that you're unfairly chastising Fred for going
for empirical evidence of success since you're doing the same thing
and, eventually, it all has to work here, in the real world.

I chastise Fred for heaping extended, crude, obscene, and frequent
insults on the majority of the posters in this group, including
newbies who ask innocent (if often naive) questions. He's not here to
learn, or to help, or to have fun, but mostly just to show how smart
he is and how stupid everybody else is. I wonder what the real story
is of Fred's life; he's silent on that.

For pete's sake, read his posts. He's rude, obscene, and perfectly
mean-spirited. If that's the sort of person you admire, well...
Click to expand...
---
I don't think I've ever said that I admire his demeanor, but I do
admire his analytical skill and ability to come up with, and
graciously give away, solutions to problems.

Also, I don't think he's _perfectly_ mean-spirited, I just think he's
got a short fuse and low tolerance for bullshit and bullshit artists.

--
John Fields
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top