Art of Electronics, 3rd edition, corrections

[Winfield Hill]

AoE 3rd edition, is shipping from Amazon, I got two copies Sunday (delivered by
the USPS). We're eager to hear about errors and typos, so we can put
corrections in the next printing, which will go to the presses next week. We'll
keep making corrections in subsequent printings, until the book is perfect. :-)

So far we have received four errors serious enough to make our errata list.
www.artofelectronics.net/errata/ The first person to find an error gets their
name up for credit (that's an error, we're not clogging the errata list with
typos, too many!).


--
Thanks,
- Win

[John Larkin]

On 7 Apr 2015 05:35:57 -0700, Winfield Hill <hi...@rowland.harvard.edu>
wrote:

Fig 3.36B Add an output

P 157 typo am plitude

P 219 some Avago phemt devices sure behave like enhancement jfets

P 265 might mention c-load opamps, like LM8261

P 294 not a mistake, but a BFT25 makes an incredible low-c, low
leakage diode

P 343 Opinion: the 34401A is a mediocre DVM, especially on AC.

P 452 Could mention coaxial ceramic resonators. And YAGs.

P 690 The zener + SCR crowbar is not robust if it has to discharge a
lot of capacitance. SCRs can fail if soft triggered, sort of a second
breakdown effect.


--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

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

[Phil Hobbs]

How many printings will Cambridge accept corrections for? Wiley cuts
them off at the 4th printing.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

[bloggs.fred...@gmail.com]

On Tuesday, April 7, 2015 at 11:41:20 AM UTC-4, John Larkin wrote:

> P 690 The zener + SCR crowbar is not robust if it has to discharge a
> lot of capacitance. SCRs can fail if soft triggered, sort of a second
> breakdown effect.

The thyristor self-triggers once IGT threshold is crossed. Underdriving IGT only makes a difference when there is not enough anode current to cause the conduction to spread across the entire device cross-section. How could that be the case with a crowbar?

http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF

[whit3rd]

On Tuesday, April 7, 2015 at 11:26:11 AM UTC-7, bloggs.fred...@gmail.com wrote:
> On Tuesday, April 7, 2015 at 11:41:20 AM UTC-4, John Larkin wrote:
>
> > P 690 The zener + SCR crowbar is not robust if it has to discharge a
> > lot of capacitance. SCRs can fail if soft triggered, sort of a second
> > breakdown effect.
>
> The thyristor self-triggers once IGT threshold is crossed. Underdriving IGT only makes a difference when there is not enough anode current to cause the conduction to spread across the entire device cross-section. How could that be the case with a crowbar?
>
> http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF

It's the dI/dt limit: the thyristor cannot deliver full current in its first microseconds, the
conduction region must spread from the gate electrode across the area of the cathode.
In a power device, that can be a few millimeters, and take a lot of microseconds.
There's fast turn-on for photoSCRs, however.

[John Larkin]

http://www.onsemi.com/pub/Collateral/MC3423-D.PDF

see sheets 5 and 6.

--

John Larkin Highland Technology, Inc

picosecond timing precision measurement

[Phil Hobbs]

On 4/7/2015 2:26 PM, bloggs.fred...@gmail.com wrote:
> On Tuesday, April 7, 2015 at 11:41:20 AM UTC-4, John Larkin wrote:
>
>> P 690 The zener + SCR crowbar is not robust if it has to discharge
>> a lot of capacitance. SCRs can fail if soft triggered, sort of a
>> second breakdown effect.
>
> The thyristor self-triggers once IGT threshold is crossed.
> Underdriving IGT only makes a difference when there is not enough
> anode current to cause the conduction to spread across the entire
> device cross-section. How could that be the case with a crowbar?
>
> http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF

It was a pretty well known problem, back when linear supplies were more
common. Special chips e.g. the MC3423 are available to fix it, by
triggering the SCR hard even as the supply voltage is collapsing.

The I**2 t handling capability of the SCR depends on the trigger
characteristics, because the triggering happens at the contact and
propagates across the die. (Centre-contact SCRs are better than
edge-contact ones for that reason, see the MC3423 datasheet.) In the
zener circuit, the SCR starves itself of gate drive as it's firing,
which slows down the propagation. The I**2 t capacity goes down
basically as the fraction of the die area that's actually fired, AIUI.

[Cursitor Doom]

Yeah, well, when you've got it as error free as the 2nd edition is, let
us know so we can all go out and buy it. :-)

[Phil Hobbs]

Too late, everybody's ordered one already. ;)

It's amazing how you can go over and over a manuscript, and then as soon
as the finished product arrives, you flip it open and start seeing
mistakes. I found about half of the errata in my first edition myself,
which took me down a peg, for sure.

[George Herold]

On Tuesday, April 7, 2015 at 5:17:20 PM UTC-4, Phil Hobbs wrote:
> On 4/7/2015 4:59 PM, Cursitor Doom wrote:
> > On Tue, 07 Apr 2015 05:35:57 -0700, Winfield Hill wrote:
> >
> >> AoE 3rd edition, is shipping from Amazon, I got two copies Sunday
> >> (delivered by the USPS). We're eager to hear about errors and typos, so
> >> we can put corrections in the next printing, which will go to the
> >> presses next week. We'll keep making corrections in subsequent
> >> printings, until the book is perfect. :-)
> >>
> >> So far we have received four errors serious enough to make our errata
> >> list. www.artofelectronics.net/errata/ The first person to find an
> >> error gets their name up for credit (that's an error, we're not clogging
> >> the errata list with typos, too many!).
> >
> > Yeah, well, when you've got it as error free as the 2nd edition is, let
> > us know so we can all go out and buy it. :-)
> >
>
> Too late, everybody's ordered one already. ;)
>
> It's amazing how you can go over and over a manuscript, and then as soon
> as the finished product arrives, you flip it open and start seeing
> mistakes. I found about half of the errata in my first edition myself,
> which took me down a peg, for sure.
>
> Cheers
>
> Phil Hobbs

I finally got notice today from amazon that mine shipped.
(the free shipping option probably puts me last in line.)
Not that I'm in any hurry, any "mistakes"
I find will likely be my own, not H&H's. :^)

George H.

[Dave Platt]

In article <5524495...@electrooptical.net>,

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

>It's amazing how you can go over and over a manuscript, and then as soon
>as the finished product arrives, you flip it open and start seeing
>mistakes. I found about half of the errata in my first edition myself,
>which took me down a peg, for sure.

It works with software too, and hardware RTL designs being sent off to
the fab, and I think a similar principle applies to interpersonal
relations as well (dropping letters in the mailbox, hitting "Send" on
the email, etc.)

There's something about taking a non-reversable step of commitment,
which manages to un-mask numerous forms of selective blindness.

Real Life is lamentably short on "Undo!" buttons which actually work.

[Tim Williams]

"Phil Hobbs" <ho...@electrooptical.net> wrote in message
news:55243E52...@electrooptical.net...

> It was a pretty well known problem, back when linear supplies were more
> common. Special chips e.g. the MC3423 are available to fix it, by
> triggering the SCR hard even as the supply voltage is collapsing.

Personally, I use a TL(V)431 with a 2N3906 common emitter inverter.
Precise threshold, accommodates hysteresis, fast yet modest gate drive.

Tim

--
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

[Winfield Hill]

Phil Hobbs wrote...

> Winfield Hill wrote:
>> We're eager to hear about errors and typos, so we can put

>> corrections in the next printing ...

>
> How many printings will Cambridge accept corrections for?
> Wiley cuts them off at the 4th printing.

I'm not aware that they have any cutoff. In the
2nd edition we stopped when corrections dropped
to nearly zero. I suspect it's pretty much up to
the dedication of our editor, which is awesome.

Cambridge University Press is awesome, in general,
because even tho they publish relatively-obscure
monographs, etc., they don't discontinue books when
the sales rate drops. But then, of course, our
sales rate never really dropped all that much.

::-))


--
Thanks,
- Win

[Winfield Hill]

John Larkin wrote...

>
> Fig 3.36B Add an output
> P 157 typo am plitude
> P 219 some Avago phemt devices sure behave like enhancement jfets
> P 265 might mention c-load opamps, like LM8261
> P 294 not a mistake, but a BFT25 makes an incredible low-c, low
> leakage diode
> P 343 Opinion: the 34401A is a mediocre DVM, especially on AC.
> P 452 Could mention coaxial ceramic resonators. And YAGs.
> P 690 The zener + SCR crowbar is not robust if it has to discharge a
> lot of capacitance. SCRs can fail if soft triggered, sort of a
> second breakdown effect.

Thanks John.


--
Thanks,
- Win

[Winfield Hill]

Tim Williams wrote...

>
>"Phil Hobbs" <ho...@electrooptical.net> wrote in message
>news:55243E52...@electrooptical.net...
>> It was a pretty well known problem, back when linear supplies were more
>> common. Special chips e.g. the MC3423 are available to fix it, by
>> triggering the SCR hard even as the supply voltage is collapsing.
>
> Personally, I use a TL(V)431 with a 2N3906 common emitter inverter.
> Precise threshold, accommodates hysteresis, fast yet modest gate drive.

How exactly do you use it? When the Ref input is below 1.24V,
what does the Cathode pin do? How low does it go (unspecified).
Low enough, below Vbe, to run an inverting transistor, EF to gnd?
We need some more details.


--
Thanks,
- Win

[Tim Williams]

"Winfield Hill" <hi...@rowland.harvard.edu> wrote in message
news:mg1v9...@drn.newsguy.com...

Example:
http://seventransistorlabs.com/Images/Crowbar2.png

The top 100 ohm sets gate current. The bottom 100 ohm sinks gate leakage
to ground, and sets positive feedback into the sense divider R1-R2. The
top 220 bleeds away bias current, and the bottom 220 prevents the TL431
from ripping the BJT apart (TL431s are rated for 100mA, though I admit I
haven't measured what they typically self-limit at -- it could be much
more?).

It's been a little while since I measured it, so I don't remember exactly
which order it's in (Vref or Va doing what), but what I remember is that a
TL431 acts something like a BJT with very well defined Vbe and very high
Gm (at DC), at the expense of a somewhat leaky sub-threshold region (so to
speak). Like, it starts turning on (uA to <1mA) in the 2.1 to 2.499V
range, then between 2.500 and 2.501V it jumps way up in current.

So, it's not simply "draws 1mA bias whether you like it or not", and then
whatever on top of that -- it does indeed have a cutoff region. It may
also be sourcing some bias from the VREF pin itself (I don't remember).

Phil mentioned "as the supply is collapsing", which is interesting; one
could "delay" the supply with a relatively large RC (a few ohms + 10s of
uF?), so the Sense Rail has something to push into the SCR. The feedback
could also be rearranged so the TL431 remains latched, rather than having
only a narrow hysteresis band.

[bloggs.fred...@gmail.com]

On Tuesday, April 7, 2015 at 4:30:14 PM UTC-4, Phil Hobbs wrote:
> On 4/7/2015 2:26 PM, bloggs.fred...@gmail.com wrote:
> > On Tuesday, April 7, 2015 at 11:41:20 AM UTC-4, John Larkin wrote:
> >
> >> P 690 The zener + SCR crowbar is not robust if it has to discharge
> >> a lot of capacitance. SCRs can fail if soft triggered, sort of a
> >> second breakdown effect.
> >
> > The thyristor self-triggers once IGT threshold is crossed.
> > Underdriving IGT only makes a difference when there is not enough
> > anode current to cause the conduction to spread across the entire
> > device cross-section. How could that be the case with a crowbar?
> >
> > http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF
>
> It was a pretty well known problem, back when linear supplies were more
> common. Special chips e.g. the MC3423 are available to fix it, by
> triggering the SCR hard even as the supply voltage is collapsing.

That chip was used mainly to obtain a precision trip threshold and transient overvoltage immunity. There is no advertisement whatsoever about driving the gate hard even as the supply is collapsing.

>
> The I**2 t handling capability of the SCR depends on the trigger
> characteristics, because the triggering happens at the contact and
> propagates across the die. (Centre-contact SCRs are better than
> edge-contact ones for that reason, see the MC3423 datasheet.) In the
> zener circuit, the SCR starves itself of gate drive as it's firing,
> which slows down the propagation. The I**2 t capacity goes down
> basically as the fraction of the die area that's actually fired, AIUI.

I2t in the case of a crowbar applies to fuse selection and not withstanding current surges. The current surge rating for a switching application is covered by ITSM, the maximum single shot switch current that can be handled without causing a damaging junction temperature rise.

The regenerative current from the anode is way more than anything you're going to inject into the gate. If the voltage across the SCR is starting collapse then you already have a pretty hefty anode current and the SCR is not likely to abruptly turn off. Larkin is always running his components outside safe operating area limits so it's no surprise to hear he was blowing up his crowbars.

[Phil Hobbs]

"Fred",

Do you actually design anything, or just rehearse your grudges on Usenet?

The zener circuit has been notorious for blowing up SCRs needlessly, for at least 30 years that I know about. That has zilch to do with personalities--yours, mine, John's, or anyone else's.

And if your views are so transparently just, why not use your real name, like a grown-up?

Cheers

Phil Hobbs

[George Herold]

On Wednesday, April 8, 2015 at 9:07:38 PM UTC-4, Phil Hobbs wrote:
> "Fred",
>
> Do you actually design anything, or just rehearse your grudges on Usenet?
>
> The zener circuit has been notorious for blowing up SCRs needlessly, for at least 30 years that I know about. That has zilch to do with personalities--yours, mine, John's, or anyone else's.

I thought it was a good sign that he at least said something
electronics related.

George H.

[Phil Hobbs]

On 4/8/2015 10:04 PM, George Herold wrote:
> On Wednesday, April 8, 2015 at 9:07:38 PM UTC-4, Phil Hobbs wrote:
>> "Fred",
>>
>> Do you actually design anything, or just rehearse your grudges on Usenet?
>>
>> The zener circuit has been notorious for blowing up SCRs needlessly, for at least 30 years that I know about. That has zilch to do with personalities--yours, mine, John's, or anyone else's.
>
> I thought it was a good sign that he at least said something
> electronics related.

Well, a better sign, anyway. ;)

[John Larkin]

On Wed, 8 Apr 2015 19:04:58 -0700 (PDT), George Herold
<ghe...@teachspin.com> wrote:

>On Wednesday, April 8, 2015 at 9:07:38 PM UTC-4, Phil Hobbs wrote:
>> "Fred",
>>
>> Do you actually design anything, or just rehearse your grudges on Usenet?
>>
>> The zener circuit has been notorious for blowing up SCRs needlessly, for at least 30 years that I know about. That has zilch to do with personalities--yours, mine, John's, or anyone else's.
>
>I thought it was a good sign that he at least said something
>electronics related.

Yes, even though he's dead wrong, and can't read a data sheet.

[John Larkin]

On Wed, 8 Apr 2015 17:17:20 -0700 (PDT),

I haven't done a crowbar circuit in decades. I do occasionally put a
tanszorb on a power rail. But soft triggering an SCR reduces its
current handling capacity.

I rarely push a part past datasheet limits; only when it seems safe
and there's a huge performance payoff.

Here's an analog computer that models junction temperature and shuts
off a mosfet when it looks too high. It's a much better way to push a
mosfet than a simple current limit, or even a foldback.

https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg

That gets combined with a heatsink temperature sensor and a time
constant to make a dynamic model of Tj.

[Bill Sloman]

On Thursday, 9 April 2015 12:54:42 UTC+10, John Larkin wrote:
> On Wed, 8 Apr 2015 19:04:58 -0700 (PDT), George Herold
> <ghe...@teachspin.com> wrote:
>
> >On Wednesday, April 8, 2015 at 9:07:38 PM UTC-4, Phil Hobbs wrote:
> >> "Fred",
> >>
> >> Do you actually design anything, or just rehearse your grudges on Usenet?
> >>
> >> The zener circuit has been notorious for blowing up SCRs needlessly, for at least 30 years that I know about. That has zilch to do with personalities--yours, mine, John's, or anyone else's.
> >
> >I thought it was a good sign that he at least said something
> >electronics related.
>
> Yes, even though he's dead wrong, and can't read a data sheet.

The fact that the data sheet doesn't mean the same thing to him as it does to John Larkin isn't definitive proof that Fred Bloggs' point of view is wrong.

Krw and John Larkin do think that way - if you can dignify the process as "thinking" - but their conclusions aren't entirely reliable.

If John had gone the trouble of pointing out what Fred Bloggs had got wrong, one might take John's post more seriously, but at this point it sounds more like yet one more squeal of injured vanity.

--
Bill Sloman, Sydney

[John S]

It applies to both. SCR specs usually include I2t values and one should
select a fuse I2t less than the SCR I2t. Or, the SCR should be selected
such that its I2t is greater than the fuse I2t.

For an example, see page 3 for I2t ratings:
<http://www.littelfuse.com/data/en/Data_Sheets/E6SCR.pdf>

[Phil Hobbs]

On 04/09/2015 12:21 AM, John Larkin wrote:

<snip>

> I rarely push a part past datasheet limits; only when it seems safe
> and there's a huge performance payoff.
>
> Here's an analog computer that models junction temperature and shuts
> off a mosfet when it looks too high. It's a much better way to push a
> mosfet than a simple current limit, or even a foldback.
>
> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg
>
> That gets combined with a heatsink temperature sensor and a time
> constant to make a dynamic model of Tj.

Neat. That's less circuitry than it would have taken to get the two
parameters into an ADC, and since the required accuracy is modest, it's
a good fit for the task.

The model is in software, I gather?

I recently did a rough square-rooter to linearize the power dissipation
in a small heater--over the relevant range, the change of slope went
from 5:1 to +-20%. That helps keep the control bandwidth constant. It
works over a pretty wide range of pass transistor temperature, too, as
long as it's got a control loop wrapped around it.

[John Larkin]

On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>On 04/09/2015 12:21 AM, John Larkin wrote:
>
><snip>
>> I rarely push a part past datasheet limits; only when it seems safe
>> and there's a huge performance payoff.
>>
>> Here's an analog computer that models junction temperature and shuts
>> off a mosfet when it looks too high. It's a much better way to push a
>> mosfet than a simple current limit, or even a foldback.
>>
>> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg
>>
>> That gets combined with a heatsink temperature sensor and a time
>> constant to make a dynamic model of Tj.
>
>Neat. That's less circuitry than it would have taken to get the two
>parameters into an ADC, and since the required accuracy is modest, it's
>a good fit for the task.

That circuit is used on a laser driver that's all analog, no uP. A
Zetex SOT23 nickel RTD picks up mosfet temp which is added to the
computed power signal, and that sum is lowpass filtered to approximate
the chip mass, then a comparator. If the customer asks for too much
power, or doesn't provide enough cooling, we shut him off.

Too bad mosfets rarely provide on-chip thermal sensing or shutdown.

>
>The model is in software, I gather?

I've done that shutdown in software too, in our NMR gradient drivers.
Digitize Ifet and the voltage across the fets, and heatsink temp, and
run a model to estimate Tj and shutdown if it looks scary. That's run
maybe 1K to 2K times per second. We tested a lot of fets to
destruction to estimate the numbers that we used in the models.

https://dl.dropboxusercontent.com/u/53724080/Parts/ExFets.jpg

>
>I recently did a rough square-rooter to linearize the power dissipation
>in a small heater--over the relevant range, the change of slope went
>from 5:1 to +-20%. That helps keep the control bandwidth constant. It
>works over a pretty wide range of pass transistor temperature, too, as
>long as it's got a control loop wrapped around it.

I like PWM drive into resistive heaters, when the system can tolerate
it. Nice and linear. One other trick is to use a mosfet as the heater;
keep the voltage mostly constant and control the current.

--

John Larkin Highland Technology, Inc

picosecond timing precision measurement

[bloggs.fred...@gmail.com]

On Wednesday, April 8, 2015 at 9:07:38 PM UTC-4, Phil Hobbs wrote:
> "Fred",
>
> Do you actually design anything, or just rehearse your grudges on Usenet?
>
> The zener circuit has been notorious for blowing up SCRs needlessly, for at least 30 years that I know about. That has zilch to do with personalities--yours, mine, John's, or anyone else's.

Those are dI/dt induced failures (hot spots due to concentrated current densities). I will concede that an unnecessarily large IGT will reduce, but not eliminate, the incidence of failure in the case of dI/dt abuse.

Care to explain how the MC3423 does anything to control dI/dt?

[George Herold]

On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
> >On 04/09/2015 12:21 AM, John Larkin wrote:
> >
> ><snip>
> >> I rarely push a part past datasheet limits; only when it seems safe
> >> and there's a huge performance payoff.
> >>
> >> Here's an analog computer that models junction temperature and shuts
> >> off a mosfet when it looks too high. It's a much better way to push a
> >> mosfet than a simple current limit, or even a foldback.
> >>
> >> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg

That is nice...
Why do you take the difference of the two collector voltages?
I've never used a diff pair, but I always see the signal
taken from one side.
Why no emitter resistors? (Would that extend linearity.. at the cost of gain?)

> >>
> >> That gets combined with a heatsink temperature sensor and a time
> >> constant to make a dynamic model of Tj.


> >
> >Neat. That's less circuitry than it would have taken to get the two
> >parameters into an ADC, and since the required accuracy is modest, it's
> >a good fit for the task.
>
> That circuit is used on a laser driver that's all analog, no uP. A
> Zetex SOT23 nickel RTD picks up mosfet temp which is added to the
> computed power signal, and that sum is lowpass filtered to approximate
> the chip mass, then a comparator. If the customer asks for too much
> power, or doesn't provide enough cooling, we shut him off.

So a slow enough low pass "time" such that the heat sink
temperature has enough time to react to the "new" power level?

>
> Too bad mosfets rarely provide on-chip thermal sensing or shutdown.
>
>
> >
> >The model is in software, I gather?
>
> I've done that shutdown in software too, in our NMR gradient drivers.
> Digitize Ifet and the voltage across the fets, and heatsink temp, and
> run a model to estimate Tj and shutdown if it looks scary. That's run
> maybe 1K to 2K times per second. We tested a lot of fets to
> destruction to estimate the numbers that we used in the models.
>
> https://dl.dropboxusercontent.com/u/53724080/Parts/ExFets.jpg
>
>
> >
> >I recently did a rough square-rooter to linearize the power dissipation
> >in a small heater--over the relevant range, the change of slope went
> >from 5:1 to +-20%. That helps keep the control bandwidth constant. It
> >works over a pretty wide range of pass transistor temperature, too, as
> >long as it's got a control loop wrapped around it.
>
> I like PWM drive into resistive heaters, when the system can tolerate
> it. Nice and linear. One other trick is to use a mosfet as the heater;
> keep the voltage mostly constant and control the current.

The mosfet heater is nice. I've also stuck both the pass transistor and
heater resistor together on the thing to be heated.

George H.

[Phil Hobbs]

Yup, in less sensitive situations I generally use PWM heating too--nice
and efficient, no thermal problems with the switch, and so on. PWM is a
no-no in this case, though, because both capacitive and inductive pickup
are serious issues. I just thought the square rooter was fun. The
layout allows me to use it or not use it. I just got the boards stuffed
by Beautiful Layout Hunchback, so we'll see how they work. (It's a
temperature control experiment that may have considerable practical
usefulness.)

[Phil Hobbs]

It doesn't, silly. What it does do is ensure enough gate drive for long
enough that the full capability of the SCR is actually available. The
Zener circuit is like a choke chain on a dog--as soon as it gets
started, it cuts off its own drive.

From the 1993 Motorola Thyristor Device Data book, p 1-5-13:

"The di/dt of the current surge pulse is also a critical parameter and
should not exceed the device's ratings.... The magnitude of di/dt that
the SCR can sustain is controlled by the device construction and, to
some extent, the gate drive conditions. When the SCR gate region is
driven on, conduction across the junction starts in a small region and
progressively propagates across the total junction. Anode current will
initially be concentrated in this small conducting area, causing high
current densities that can degrade and ultimately destroy the device.
*To minimize this di/dt effect, the gate should be turned on hard and
fast such that the area turned on is initially maximized.* This can be
accomplished with a gate current pulse approaching five times the
maximum continuous gate current I_gt, and with a fast rise time (< 1
us). The gate current pulse width should be greater than the
propagation time; a figure of 10 us minimum should satisfy most SCRs
with average current ratings under 50A or so."

(emphasis added)

Official enough for you?

Like I said, this problem has been known forever.

[John Larkin]

On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
<ghe...@teachspin.com> wrote:

>On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
>> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
>> <pcdhSpamM...@electrooptical.net> wrote:
>>
>> >On 04/09/2015 12:21 AM, John Larkin wrote:
>> >
>> ><snip>
>> >> I rarely push a part past datasheet limits; only when it seems safe
>> >> and there's a huge performance payoff.
>> >>
>> >> Here's an analog computer that models junction temperature and shuts
>> >> off a mosfet when it looks too high. It's a much better way to push a
>> >> mosfet than a simple current limit, or even a foldback.
>> >>
>> >> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg
>
>That is nice...
>Why do you take the difference of the two collector voltages?

That's in fact the signal that we want, approximately the product of
Vfet*Ifet.

>I've never used a diff pair, but I always see the signal
>taken from one side.
>Why no emitter resistors? (Would that extend linearity.. at the cost of gain?)

We don't want linearity! It's a multiplier!

[Phil Hobbs]

On 04/09/2015 03:11 PM, John Larkin wrote:
> On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
> <ghe...@teachspin.com> wrote:
>
>> On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
>>> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
>>> <pcdhSpamM...@electrooptical.net> wrote:
>>>
>>>> On 04/09/2015 12:21 AM, John Larkin wrote:
>>>>
>>>> <snip>
>>>>> I rarely push a part past datasheet limits; only when it seems safe
>>>>> and there's a huge performance payoff.
>>>>>
>>>>> Here's an analog computer that models junction temperature and shuts
>>>>> off a mosfet when it looks too high. It's a much better way to push a
>>>>> mosfet than a simple current limit, or even a foldback.
>>>>>
>>>>> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg
>>
>> That is nice...
>> Why do you take the difference of the two collector voltages?
>
> That's in fact the signal that we want, approximately the product of
> Vfet*Ifet.

The difference voltage goes as tanh(delta Vbe*e/2kT), which has only odd
order terms, so the even order nonlinearity in the individual collector
currents cancels.

>
>> I've never used a diff pair, but I always see the signal
>> taken from one side.
>> Why no emitter resistors? (Would that extend linearity.. at the cost of gain?)
>
> We don't want linearity! It's a multiplier!

Another approach would be an LM13700, but there are a lot more dual
transistors than OTAs in the world, and they're cheaper and smaller, to
boot. (I'm a big OTA fan in principle, having grown up on NS linear
applications manuals, but I've only ever used them in one-offs.

[George Herold]

On Thursday, April 9, 2015 at 3:12:04 PM UTC-4, John Larkin wrote:
> On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
> <ghe...@teachspin.com> wrote:
>
> >On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
> >> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
> >> <pcdhSpamM...@electrooptical.net> wrote:
> >>
> >> >On 04/09/2015 12:21 AM, John Larkin wrote:
> >> >
> >> ><snip>
> >> >> I rarely push a part past datasheet limits; only when it seems safe
> >> >> and there's a huge performance payoff.
> >> >>
> >> >> Here's an analog computer that models junction temperature and shuts
> >> >> off a mosfet when it looks too high. It's a much better way to push a
> >> >> mosfet than a simple current limit, or even a foldback.
> >> >>
> >> >> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg
> >
> >That is nice...
> >Why do you take the difference of the two collector voltages?
>
> That's in fact the signal that we want, approximately the product of
> Vfet*Ifet.
>
> >I've never used a diff pair, but I always see the signal
> >taken from one side.
> >Why no emitter resistors? (Would that extend linearity.. at the cost of gain?)
>
> We don't want linearity! It's a multiplier!

Ouch, you're yelling. :^)
Remember, I am but an egg in the sea of electronics.

Sorry, I'm at home with (appropriately) my first ed. of AoE.
Without the emitter resisters, the gain goes as the
base-emitter resistance, r-sub-e in AoE notation.
Which is linear with the collector current.

I got it..

Re: diff opamp. The signal in each side is equal and opposite
You could just take the signal from one side, right?
I hope it's OK I ask questions?

George H.

[Jim Thompson]

On Thu, 9 Apr 2015 16:12:53 -0700 (PDT), George Herold

<ghe...@teachspin.com> wrote:

>On Thursday, April 9, 2015 at 3:12:04 PM UTC-4, John Larkin wrote:
>> On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
>> <ghe...@teachspin.com> wrote:
>>
>> >On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
>> >> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
>> >> <pcdhSpamM...@electrooptical.net> wrote:

[snip]

>>
>> We don't want linearity! It's a multiplier!
>Ouch, you're yelling. :^)
>Remember, I am but an egg in the sea of electronics.
>
>Sorry, I'm at home with (appropriately) my first ed. of AoE.
>Without the emitter resisters, the gain goes as the
>base-emitter resistance, r-sub-e in AoE notation.
>Which is linear with the collector current.
>
>I got it..
>

[snip]

I don't think quite. You need to write out the I-V equations for the
B-E junctions, the effective TANH transfer function is what yields the
multiplication.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

[George Herold]

On Thursday, April 9, 2015 at 3:29:42 PM UTC-4, Phil Hobbs wrote:
> On 04/09/2015 03:11 PM, John Larkin wrote:
> > On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
> > <ghe...@teachspin.com> wrote:
> >
> >> On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
> >>> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
> >>> <pcdhSpamM...@electrooptical.net> wrote:
> >>>
> >>>> On 04/09/2015 12:21 AM, John Larkin wrote:
> >>>>
> >>>> <snip>
> >>>>> I rarely push a part past datasheet limits; only when it seems safe
> >>>>> and there's a huge performance payoff.
> >>>>>
> >>>>> Here's an analog computer that models junction temperature and shuts
> >>>>> off a mosfet when it looks too high. It's a much better way to push a
> >>>>> mosfet than a simple current limit, or even a foldback.
> >>>>>
> >>>>> https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg
> >>
> >> That is nice...
> >> Why do you take the difference of the two collector voltages?
> >
> > That's in fact the signal that we want, approximately the product of
> > Vfet*Ifet.
>
> The difference voltage goes as tanh(delta Vbe*e/2kT), which has only odd
> order terms, so the even order nonlinearity in the individual collector
> currents cancels.

Wow... I think I understand that. Tanks.

>
> >
> >> I've never used a diff pair, but I always see the signal
> >> taken from one side.
> >> Why no emitter resistors? (Would that extend linearity.. at the cost of gain?)
> >
> > We don't want linearity! It's a multiplier!
>
> Another approach would be an LM13700, but there are a lot more dual
> transistors than OTAs in the world, and they're cheaper and smaller, to
> boot. (I'm a big OTA fan in principle, having grown up on NS linear
> applications manuals, but I've only ever used them in one-offs.

OK.. I've heard OTA's can be used as multipliers,
but I need to do more homework... it's again got
gain control that goes as some current. (right?)
I've only used expensive multipliers from AD $5-15.
or mixers from minicircuits.

George H.

[George Herold]

On Thursday, April 9, 2015 at 7:18:55 PM UTC-4, Jim Thompson wrote:
> On Thu, 9 Apr 2015 16:12:53 -0700 (PDT), George Herold
> <ghe...@teachspin.com> wrote:
>
> >On Thursday, April 9, 2015 at 3:12:04 PM UTC-4, John Larkin wrote:
> >> On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
> >> <ghe...@teachspin.com> wrote:
> >>
> >> >On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
> >> >> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
> >> >> <pcdhSpamM...@electrooptical.net> wrote:
> [snip]
> >>
> >> We don't want linearity! It's a multiplier!
> >Ouch, you're yelling. :^)
> >Remember, I am but an egg in the sea of electronics.
> >
> >Sorry, I'm at home with (appropriately) my first ed. of AoE.
> >Without the emitter resisters, the gain goes as the
> >base-emitter resistance, r-sub-e in AoE notation.
> >Which is linear with the collector current.
> >
> >I got it..
> >
> [snip]
>
> I don't think quite. You need to write out the I-V equations for the
> B-E junctions, the effective TANH transfer function is what yields the
> multiplication.

OK more homework! It's warm ~60's, but raining, here now.

I assume I start with the Eber's Moll model for re vs Ic.

George H.

[Jim Thompson]

On Thu, 9 Apr 2015 16:28:52 -0700 (PDT), George Herold

<ghe...@teachspin.com> wrote:

>On Thursday, April 9, 2015 at 7:18:55 PM UTC-4, Jim Thompson wrote:
>> On Thu, 9 Apr 2015 16:12:53 -0700 (PDT), George Herold
>> <ghe...@teachspin.com> wrote:
>>
>> >On Thursday, April 9, 2015 at 3:12:04 PM UTC-4, John Larkin wrote:
>> >> On Thu, 9 Apr 2015 11:39:13 -0700 (PDT), George Herold
>> >> <ghe...@teachspin.com> wrote:
>> >>
>> >> >On Thursday, April 9, 2015 at 1:46:33 PM UTC-4, John Larkin wrote:
>> >> >> On Thu, 09 Apr 2015 12:34:06 -0400, Phil Hobbs
>> >> >> <pcdhSpamM...@electrooptical.net> wrote:
>> [snip]
>> >>
>> >> We don't want linearity! It's a multiplier!
>> >Ouch, you're yelling. :^)
>> >Remember, I am but an egg in the sea of electronics.
>> >
>> >Sorry, I'm at home with (appropriately) my first ed. of AoE.
>> >Without the emitter resisters, the gain goes as the
>> >base-emitter resistance, r-sub-e in AoE notation.
>> >Which is linear with the collector current.
>> >
>> >I got it..
>> >
>> [snip]
>>
>> I don't think quite. You need to write out the I-V equations for the
>> B-E junctions, the effective TANH transfer function is what yields the
>> multiplication.
>
>OK more homework! It's warm ~60's, but raining, here now.

79°F here ;-) Time to grab a glass of wine and head out to the patio
to watch O'Reilly >:-}

>
>I assume I start with the Eber's Moll model for re vs Ic.
>
>George H.
>

Yes. Sort of. Stop thinking _resistance_.

[Bill Sloman]

It doesn't have to be.

In my millidegree temperature controller, we use PWM to control the current running into the Peltier junction. The PWM is actually set up as a 10-bit binary number, and the microprocessor that sets the number takes into account the temperature difference across the Peltier junction and gives it a lot more current when it's cooling hard than when it's acting as a heater.

The response of the junction - in terms of watts transferred per amp of current through the junction - varies by 7:1 over the original design temperature range.

Sloman A.W., Buggs P., Molloy J., and Stewart D. "A microcontroller-based driver to stabilise the temperature of an optical stage to 1mK in the range 4C to 38C, using a Peltier heat pump and a thermistor sensor" Measurement Science and Technology, 7 1653-64 (1996).

This makes the transfer function distinctly non-linear, and the microprocessor linearises it for us, allowing us to keep the loop close to critically damped.

The integrated circuit Peltier controllers you can buy from Intel and Linear Technology are linear, and can't play that particular trick.

Sloman A.W. "Comment on 'Implementing of a precision fast thermoelectric cooler controller using a personal computer parallel port connection and ADV8830 controller'[Rev.Sci. Instrum. 74, 3862 (2003)]" Review of Scientific Instruments, 75 788-9 (2004)."

--
Bill Sloman, Sydney

[George Herold]

Hmm.. follow the current? I must admit at some level
bjt's are still a mystery. (Fet's are easier to understand)

My wife has been a total trooper today, and my must important job
right now is to go take care of the kitchen.
Tomorrow,
George H.

[Phil Hobbs]

An OTA is a diff pair and a bunch of current mirrors. The bias current
(I_ABC in the LM13700 datasheet, for Amplifier Bias Current) is mirrored
to become the tail current for the input differential pair. It's really
just a glorified version of the circuit John L posted.

The output is basically the collector currents of the input pair,
mirrored so that one comes from the positive rail and the other from the
negative rail. So all the level shifting is done for you, and you can
do whatever you like with the difference current.

They're on the noisy side, due to getting about 3 copies of full shot
noise on each current from all the BJT pairs.

The LM13700 also has the Gilbert feature, i.e. diode-connected
transistors on each input that you can use to make it into a linear
multiplier.

[John Larkin]

On Thu, 9 Apr 2015 16:12:53 -0700 (PDT), George Herold

No. We're trying to compute the product VMON*IMON.

Imagine that the mosfet current indicator VMON is zero, but the drain
voltage VMON varies. Both PNP bases are at 0 volts, and the current
through R1 (which is symbolic of fet D-S voltage VMON) splits between
the two PNPs. Both voltages Z1 and Z2 go up as VMON goes up, but
there's no power dissipation in the fet, and we're trying to compute
dissipation.

The *difference* between the collector currents represents mosfet
power.

!!!

>I hope it's OK I ask questions?

Sure. Lots of people don't like this circuit, especially returning the
PNP bases and collectors to ground. Too bad, it works.

--

John Larkin Highland Technology, Inc

picosecond timing laser drivers and controllers

[John Larkin]

oops ^^^^ IMON

[Jim Thompson]

On Thu, 9 Apr 2015 17:32:48 -0700 (PDT), George Herold

I totally understand.

[Phil Hobbs]

Far too many elementary electronics texts state that in order for a BJT
to be in normal bias, the CB junction has to be _reverse-biased_.

That's actually true if you account for the built-in junction potential,
but not if you're measuring voltages in the external circuit. Good BJTs
will operate nearly normally with V_CE ~ 0.3V or even a bit less. They
do slow down on account of the increased capacitances.

Current mirrors are full of devices where V_CE = V_CB ~ 0.6V.

[bloggs.fred...@gmail.com]

No- it's just a bunch of junk written by an intern about 40 year old technology.
If you read the HDK in detail, you will find the dI/dt is spec'd for "recommended" gate drives, which are anything above the maximum minimum gate trigger level. Of course, it's understood that once the gate is triggered, the drive is applied long enough for complete turn-on.

It's going to be hard to destroy something more modern like this (Fig.10):

http://www.littelfuse.com/~/media/electronics/datasheets/switching_thyristors/littelfuse_thyristor_sxx55x_datasheet.pdf.pdf

Those old zener circuits predated the precision threshold gate drives but they're still usable for low energy discharges. You can use them to discharge the massively large energy storage circuits by going with something like this:

Please view in a fixed-width font such as Courier.

.
.
.
.
. ----+-------------------------------+----
. | |
. | |
. | --------|<|-------------+
. | | D |
. ---/ | |
. // \ DZ | [Z]
. --- | |
. | | |
. | | |
. ------|--------------- ---
. | | \ / SCR2
. | SCR1 | ---
. +--[RL]--|>|--------+---/|
. | \ | | |
. | | | | |
. | +-- | |
. |0.1u | | |
. === [Rz] [Rg] |
. | | | |
. -----------+------------+------+----+----
.
.
.
. SCR1 small
.
. SCR2 large
.
.
.

[Phil Hobbs]

Whereas SCRs are so new, and all. ;)

We were discussing the hoary age of the problem, after all. And if you
have to make up lies about its author to discredit it, I think my point
is adequately made.

If you read the HDK in detail, you will find the dI/dt is
> spec'd for "recommended" gate drives, which are anything above the
> maximum minimum gate trigger level. Of course, it's understood that
> once the gate is triggered, the drive is applied long enough for
> complete turn-on.

But the zener circuit doesn't always do that, especially with slowly
varying inputs. Something with some snap helps. The zener circuit
isn't guaranteed to destroy the SCR, but as the databook says, using
something just a bit better makes sure you get the rated capacity of the
SCR. Your circuit looks like it would work fine, but it has a lot of
parts compared with, for instance, using a 723 or 431 type voltage
regulator chip. (The MC3423 has been discontinued.)

Like I said, the problem is simple to understand, easy to fix, and has
been known for decades.

[bloggs.fred...@gmail.com]

Actually they are...do you think stuff like the azipod 17MW 6600V synchronous motor cycloconverters are built from 1960s thyristor technology?

>
> We were discussing the hoary age of the problem, after all. And if you
> have to make up lies about its author to discredit it, I think my point
> is adequately made.

The particular section you quoted was redacted by a marketeer.

>
> If you read the HDK in detail, you will find the dI/dt is
> > spec'd for "recommended" gate drives, which are anything above the
> > maximum minimum gate trigger level. Of course, it's understood that
> > once the gate is triggered, the drive is applied long enough for
> > complete turn-on.
>
> But the zener circuit doesn't always do that, especially with slowly
> varying inputs. Something with some snap helps.

The thyristor has all the "snap" you need. Once the main terminal currents hit a threshold where the sum of junction current gains ( current dependent alphas) exceed one, it takes off on its own, that threshold should occur at spec'd IGT, and that abusive gate drive of 5x abs max they quoted isn't going to make things switch much faster.

[John Larkin]

What a horrible data sheet. No definitions, few contexts, no test
circuits shown. Fig 10 doesn't mention the gate drive.

I'd expect that you could destroy it with a simple zener crowbar
circuit. I certainly wouldn't assume that a soft trigger is safe.

"Modern" chips tend to be physically small, which can make them
cheaper and less rugged.

--

John Larkin Highland Technology, Inc

picosecond timing precision measurement

[bloggs.fred...@gmail.com]

Right, TECCOR doesn't know what they're doing...

[John Larkin]

On Fri, 10 Apr 2015 13:48:11 -0700 (PDT),

Whoever writes their data sheets doesn't.

[mikk...@gmail.com]

On Friday, April 10, 2015 at 5:49:22 AM UTC+3, Phil Hobbs wrote:
> Far too many elementary electronics texts state that in order for a BJT
> to be in normal bias, the CB junction has to be _reverse-biased_.
>
> That's actually true if you account for the built-in junction potential,
> but not if you're measuring voltages in the external circuit. Good BJTs
> will operate nearly normally with V_CE ~ 0.3V or even a bit less. They
> do slow down on account of the increased capacitances.
>
> Current mirrors are full of devices where V_CE = V_CB ~ 0.6V.
>
> Cheers
>
> Phil Hobbs

SiGe's and and in particular cryogenic SiGe's are agile to this:
the B-E drop increases to 0.9 - 1.1 V ballpark and you can pull the
collector significantly below the base without ... well, I would
hesitate to say 'any' ... er ... noticeable adverse effect.

There's the neat circuit doi:10.1109/TASC.2012.2227638
by Drung utilizing this:

Vsupply

I
.--o--.
I I
.----o o---.
I I I I
I I I I
I I/ \I I
Vin+ ------I I------ Vin-
I I> <I I
I I I I
I I I I
I I I I
I I/ \I I
\--I I--/
I> <I
I I
I I
v v

Io+ Io-
To room temperature

Regards,
Mikko

[Winfield Hill]

Winfield Hill wrote...
>
> AoE 3rd edition ... We're eager to hear about errors and typos,
> so we can put corrections in the next printing ...

The errors are beginning to stream in now, see our errata page,
http://artofelectronics.net/errata/ The 1st printing of 15k
books has, or is, running out. Yesterday our editor forwarded
a set of corrections to apply before the 2nd printing. If any
more come in today, maybe they can be squeezed under the wire.



--
Thanks,
- Win

[piglet]

On 11/04/2015 14:15, Winfield Hill wrote:
> Winfield Hill wrote...
>>
>> AoE 3rd edition ... We're eager to hear about errors and typos,
>> so we can put corrections in the next printing ...
>

I keep seeing references to chapter 3x - but cannot find any chapter 3x.
For instance p170 col 2 line 15. "Going beyond this first look at jfets,
chapter 3x includes some ..." or for instance last line of p167 where
reader is referred to para 3x.4

Is the "X" a placeholder that was going to be filled-in?

piglet

[Winfield Hill]

piglet wrote...

As we explain in the Preface (fittingly on page xxx), a
companion book, "The Art of Electronics: The x-Chapters"
will be coming out in a few years. This book will have
piles of advanced material we couldn't fit into the 3rd
edition. For the portions we've already written, we can
give you some cross-references now. These are indicated
with an x after the section number, etc., for brevity.
You'll find these x-Chapter cross references all over,
but IIRC, especially in Chapter 3.


--
Thanks,
- Win

[Phil Hobbs]

Cute. That's a lot like the two-JFET cascode trick, with a higher-Vp
part on top and a lower-Vp one on the bottom.

[piglet]

On 11/04/2015 17:11, Winfield Hill wrote:
> As we explain in the Preface (fittingly on page xxx), a
> companion book, "The Art of Electronics: The x-Chapters"
> will be coming out in a few years. This book will have

Ahh, that explains it! Thanks Win, I was so excited when the book
arrived that I just dived straight in without reading the preface.
Waiting several years is going to be tough but worth it.

piglet

[Chris Jones]

[snip]

> Another approach would be an LM13700, but there are a lot more dual
> transistors than OTAs in the world, and they're cheaper and smaller, to
> boot. (I'm a big OTA fan in principle, having grown up on NS linear
> applications manuals, but I've only ever used them in one-offs.
>
> Cheers
>
> Phil Hobbs
>
>

I think there is a chicken and egg problem here - few people buy OTAs so
the marketing departments at chip companies don't know that they are
nice things, so none of them have been designed for decades, so there
are none of them with rail to rail outputs or modern processes etc. so
few people buy them.

They are used a lot within chips - perhaps as much as opamps.

My favourite use is for driving the gates of large MOSFETs in things
like constant current sinks. The gate capacitance helps the compensation
and it is stable, whereas if one uses an op-amp then the op-amp output
impedance and the gate capacitance forms a pole, and along with the
other pole in the op-amp compensation, and any parasitic poles it is
very likely to oscillate.

Chris

[Phil Hobbs]

On 4/13/2015 2:32 AM, Chris Jones wrote:
> [snip]
>> Another approach would be an LM13700, but there are a lot more dual
>> transistors than OTAs in the world, and they're cheaper and smaller, to
>> boot. (I'm a big OTA fan in principle, having grown up on NS linear
>> applications manuals, but I've only ever used them in one-offs.
>>
>> Cheers
>>
>> Phil Hobbs
>>
>>
>
> I think there is a chicken and egg problem here - few people buy OTAs so
> the marketing departments at chip companies don't know that they are
> nice things, so none of them have been designed for decades, so there
> are none of them with rail to rail outputs or modern processes etc. so
> few people buy them.

The LM13700 is pretty nearly rail-to-rail, if you don't use the buffer.
My main issue is that it's too slow for almost anything--only about 1
MHz at maximum I_ABC, and slower still at lower current. I have a
couple of dozen VTC VA713 and VA2713 70-MHz OTAs in my drawer from 20
years ago, but they're long obsolete.

>
> They are used a lot within chips - perhaps as much as opamps.

Bits of them, for sure--diff pairs driving current mirrors for level
shifting. Dunno about full OTAs, but maybe so.

>
> My favourite use is for driving the gates of large MOSFETs in things
> like constant current sinks. The gate capacitance helps the compensation
> and it is stable, whereas if one uses an op-amp then the op-amp output
> impedance and the gate capacitance forms a pole, and along with the
> other pole in the op-amp compensation, and any parasitic poles it is
> very likely to oscillate.

That's an interesting approach--put the dominant pole on the output.
Parts like the LM8261 do that as well.

[Jim Thompson]

What characteristics are desired in an OTA? Defined gm? Dominant
pole at output?

[John Larkin]

On Mon, 13 Apr 2015 10:13:10 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

Lots of RRO amps are c-load stable. AD8565 for example.

--

John Larkin Highland Technology, Inc

picosecond timing laser drivers and controllers

[Phil Hobbs]

>What characteristics are desired in an OTA?  Defined gm?  Dominant
pole at output?

Differential inputs, low output distortion, low offset, optional linearizing diodes as in the 13700, and much higher speed, say >10 MHz at 100 uA tail current.

TI makes a weird part (OPA860) that they call an OTA but it's more of an improved BJT.

Cheers

Phil Hobbs

[John Larkin]

On 7 Apr 2015 05:35:57 -0700, Winfield Hill <hi...@rowland.harvard.edu>
wrote:

>AoE 3rd edition, is shipping from Amazon, I got two copies Sunday (delivered by
>the USPS). We're eager to hear about errors and typos, so we can put
>corrections in the next printing, which will go to the presses next week. We'll
>keep making corrections in subsequent printings, until the book is perfect. :-)
>
>So far we have received four errors serious enough to make our errata list.
>www.artofelectronics.net/errata/ The first person to find an error gets their
>name up for credit (that's an error, we're not clogging the errata list with
>typos, too many!).

Fig 9.113 is rather hard on the pot.

[Winfield Hill]

John Larkin wrote...

>
> Fig 9.113 is rather hard on the pot.

Yes. There's an interesting story about this. My experience
goes back to the circuit suggested by Harris Semiconductor,
for their HIP5600 high-voltage regulator. It uses the same
scheme, an LM317-type voltage-setting resistor to ground.
The pot would be an 2-watt Allen Bradley type J, or similar.
When you turn down the voltage, the pot must discharge the
output capacitor. However, it seems to be perfectly happy
enough to doing this.


--
Thanks,
- Win

[Bill Beaty]

On Thursday, April 9, 2015 at 5:32:53 PM UTC-7, George Herold wrote:
> On Thursday, April 9, 2015 at 7:50:49 PM UTC-4, Jim Thompson wrote:
> > Yes. Sort of. Stop thinking _resistance_.
> Hmm.. follow the current? I must admit at some level

> bjt's are still a mystery.

> (Fet's are easier to understand)

...except when trying to figure out what actually happens inside the channel during pinchoff. That place is weird. Shouldn't it go unstable and start spalling off moving blobs of depletion, even spewing out millimeter waves like a tiny little TWT amp? And I recall an IR microscope photo of a horizontal mosfet channel emitting light as part of the usual breakdown regime. "Usual?" The FET channel in pinchoff is constantly breaking down? Yup, apparently.

On understanding BJTs, I can't resist here pushing my beginners' non-math article on Ebers-Moll which doesn't discuss Ebers-Moll:

http://amasci.com/amateur/transis.html


---
(((((((((((((((( ( ( ( ( (O) ) ) ) ) ))))))))))))))))
William J. Beaty http://staff.washington.edu/wbeaty/
beaty, chem washington edu Research Engineer
billb, amasci com UW Chem Dept, BAG74
3-6195 Box 351700, Seattle, WA 98195-1700

[Bill Beaty]

On Saturday, April 11, 2015 at 6:16:02 AM UTC-7, Winfield Hill wrote:
> The errors are beginning to stream in now, see our errata page,

Hey, someone from cambridge.org contacted me about spreading the word on AOE-3, but then they never sent me any ads to put on amasci.com. Were none made?

Found it, Feb 3 Fran Robinson, no email response.

(((((((((((((((( ( ( ( ( (O) ) ) ) ) ))))))))))))))))
William J. Beaty http://staff.washington.edu/wbeaty/
beaty, chem washington edu Research Engineer

billb, amasci com UW Chem Dept, Bagley Hall RM74
x3-6195 Box 351700, Seattle, WA 98195-1700

[George Herold]

On Tuesday, April 14, 2015 at 9:21:47 PM UTC-4, Bill Beaty wrote:
> On Thursday, April 9, 2015 at 5:32:53 PM UTC-7, George Herold wrote:
> > On Thursday, April 9, 2015 at 7:50:49 PM UTC-4, Jim Thompson wrote:
> > > Yes. Sort of. Stop thinking _resistance_.
> > Hmm.. follow the current? I must admit at some level
>
> > bjt's are still a mystery.
>
> > (Fet's are easier to understand)
>
> ...except when trying to figure out what actually happens inside the channel during pinchoff. That place is weird. Shouldn't it go unstable and start spalling off moving blobs of depletion, even spewing out millimeter waves like a tiny little TWT amp? And I recall an IR microscope photo of a horizontal mosfet channel emitting light as part of the usual breakdown regime. "Usual?" The FET channel in pinchoff is constantly breaking down? Yup, apparently.
>
> On understanding BJTs, I can't resist here pushing my beginners' non-math article on Ebers-Moll which doesn't discuss Ebers-Moll:
>
> http://amasci.com/amateur/transis.html

Hi Bill, I read a page or so... I didn't get to bjt's

You know I said this the other day at work,
and a smart electronics guy responded,
"One elctron/hole from E to B base drags more along with it
that fall into the collector".

And I thought, "No, it's Vbe that matter's.
I have to pull some charge carriers out of the
base to get Vbe lower.
And that reduces the drift current,
and so the carriers diffuse further.

It's more diffusion that gives more current.
They are not pulled, they diffuse.

At least that's my current model.
(Does that work for you?)

George H.

[John Larkin]

On Tue, 14 Apr 2015 18:21:42 -0700 (PDT), Bill Beaty
<bi...@eskimo.com> wrote:

>On Thursday, April 9, 2015 at 5:32:53 PM UTC-7, George Herold wrote:
>> On Thursday, April 9, 2015 at 7:50:49 PM UTC-4, Jim Thompson wrote:
>> > Yes. Sort of. Stop thinking _resistance_.
>> Hmm.. follow the current? I must admit at some level
>
>> bjt's are still a mystery.
>
>> (Fet's are easier to understand)
>
>...except when trying to figure out what actually happens inside the channel during pinchoff. That place is weird. Shouldn't it go unstable and start spalling off moving blobs of depletion, even spewing out millimeter waves like a tiny little TWT amp? And I recall an IR microscope photo of a horizontal mosfet channel emitting light as part of the usual breakdown regime. "Usual?" The FET channel in pinchoff is constantly breaking down? Yup, apparently.
>
>On understanding BJTs, I can't resist here pushing my beginners' non-math article on Ebers-Moll which doesn't discuss Ebers-Moll:
>
>http://amasci.com/amateur/transis.html
>
>

I don't understand how diodes or BJTs or fets work. Doesn't seem to
matter. I do understand pretty much how they behave.

[Winfield Hill]

Bill Beaty wrote...

Somebody in marketing, no doubt. Dunno, haven't seen any ads.


--
Thanks,
- Win

[Bill Beaty]

On Wednesday, April 15, 2015 at 4:18:37 AM UTC-7, Winfield Hill wrote:
>
> Somebody in marketing, no doubt. Dunno, haven't seen any ads.

I'll try emailing again.

We should all plaster some AOE banner ads all over our blogs!

[Winfield Hill]

Phil Hobbs wrote...

> John Larkin wrote:
>> Phil Hobbs wrote:

>>> On 04/09/2015 12:21 AM, John Larkin wrote:
>>>
>>> <snip>
>>>> I rarely push a part past datasheet limits; only when it seems safe
>>>> and there's a huge performance payoff.
>>>>
>>>> Here's an analog computer that models junction temperature and shuts
>>>> off a mosfet when it looks too high. It's a much better way to push a
>>>> mosfet than a simple current limit, or even a foldback.
>>>>
>>>>
<https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/Soar_Calc.jpg>
>>>>
>>>> That gets combined with a heatsink temperature sensor and a time
>>>> constant to make a dynamic model of Tj.
>>>
>>> Neat. That's less circuitry than it would have taken to get the two
>>> parameters into an ADC, and since the required accuracy is modest, it's
>>> a good fit for the task.
>>
>> That circuit is used on a laser driver that's all analog, no uP. A
>> Zetex SOT23 nickel RTD picks up mosfet temp which is added to the
>> computed power signal, and that sum is lowpass filtered to approximate
>> the chip mass, then a comparator. If the customer asks for too much
>> power, or doesn't provide enough cooling, we shut him off.
>>
>> Too bad mosfets rarely provide on-chip thermal sensing or shutdown.
>>
>>> The model is in software, I gather?
>>
>> I've done that shutdown in software too, in our NMR gradient drivers.
>> Digitize Ifet and the voltage across the fets, and heatsink temp, and
>> run a model to estimate Tj and shutdown if it looks scary. That's run
>> maybe 1K to 2K times per second. We tested a lot of fets to
>> destruction to estimate the numbers that we used in the models.
>>
>> https://dl.dropboxusercontent.com/u/53724080/Parts/ExFets.jpg
>>
>>> I recently did a rough square-rooter to linearize the power dissipation
>>> in a small heater--over the relevant range, the change of slope went
>>>from 5:1 to +-20%. That helps keep the control bandwidth constant. It
>>> works over a pretty wide range of pass transistor temperature, too, as
>>> long as it's got a control loop wrapped around it.
>>
>> I like PWM drive into resistive heaters, when the system can tolerate
>> it. Nice and linear. One other trick is to use a mosfet as the heater;
>> keep the voltage mostly constant and control the current.

Yes, a square root in the feedback loop is called for if the
integrated error signal controls both the raw supply voltage,
and the PWM output. In such a case, a 50:1 output control
range can easily be extended to a 2500:1 dynamic range.

> Yup, in less sensitive situations I generally use PWM heating too--nice
> and efficient, no thermal problems with the switch, and so on. PWM is
> a no-no in this case, though, because both capacitive and inductive
> pickup are serious issues. I just thought the square rooter was fun.
> The layout allows me to use it or not use it. I just got the boards
> stuffed by Beautiful Layout Hunchback, so we'll see how they work.
> (It's a temperature control experiment that may have considerable
> practical usefulness.)

Tell us about Beautiful Layout Hunchback.


--
Thanks,
- Win

[Phil Hobbs]

Two of my children have been helping out with various things, and we (or
at least I) have a running joke about it being hard to get good
hunchbacks these days.

Dashing Firmware Hunchback is my son Simon, who just finished his third
year in Honours Physics at UBC. He spent two summers working with me
building gizmos, which will be a leg up when he goes off to grad school.
This summer he's going to be working at the local mini particle
accelerator/laboratory astrophysics lab (TRIUMF) on a dark-matter
detection experiment, where it sounds like he gets to be the main gizmo
guy. (He'll be doing a bunch of FPGAs and other circuitry to look for
correlations from 256 photomultipliers. He's really jazzed about it.)

Beautiful Layout Hunchback is my younger daughter, Magdalen, who
recently graduated with a liberal arts degree. To fill in till the job
of her dreams comes along, she's been doing PC board layouts, board
stuffing, inventory, all that hunchback-type stuff, and studying physics
and math part time (which she discovers that she loves).

It's a wonderful thing to be able to admire one's own children.

[John Larkin]

I used to use Liz, aka The Brat, for PCB layout. But she went and
hired a layout guy, and signed up for an MBA, so I guess she's a
manager now. Looks like I'll be working for her soon too. Suits me
fine.

[Don Kuenz]

The project on the cover of the January/February 2016 edition of _QEX_
uses digitally tunable capacitors. The good book, AoE, mentions digital
capacitors on page 1082. Once again, the enigmatic Chapter 3x appears.
;)

"My world is coming nearer."
--
Don Kuenz KB7RPU

[John Larkin]

I've used the 3-step Maxim "Flecap" as the coarse tune of a VCO, with
a varicap for the fine tune. Worked fine. Other people make apparently
better digital caps now.

Maxim has EOL'd it of course. Never Buy Maxim.

--

John Larkin Highland Technology, Inc

picosecond timing precision measurement

[John Larkin]

32 step, actually.

[Don Kuenz]

The _QEX_ project uses a 32 step Peregrine PE64102 [1].

Note.

1. http://www.psemi.com/pdf/datasheets/pe64102ds.pdf

--
Don Kuenz KB7RPU

[John Larkin]

On Thu, 7 Jan 2016 22:52:34 -0000 (UTC), Don Kuenz

That part has a specified operating frequency range of 100M to 3G Hz.
I wonder why there is a minimum.

A couple of other peaple make digital caps, but they seem to use MEMS
technology. That's usually bad news.

[Winfield Hill]

John Larkin wrote...
> John Larkin wrote:
>
>> I've used the 32-step Maxim "Flecap" as the coarse tune

>> of a VCO, with a varicap for the fine tune. Worked fine.
>> Other people make apparently better digital caps now.
>>
>> Maxim has EOL'd it of course. Never Buy Maxim.
>
> 32 step, actually.

Get a lifetime supply, 6k at Mouser
and 13k at Rochester, 75 cents each.


--
Thanks,
- Win

[John Larkin]

On 9 Jan 2016 09:44:08 -0800, Winfield Hill <hi...@rowland.harvard.edu>
wrote:

I'd better. I don't want to spin the board layout or, even worse,
rewrite the code, which is 68K assembly.

I'm designing a new oscillator now, and prefer to use a small
surface-mount trimmer cap to center the frequency. One could tweak
that with a screwdriver in 30 seconds. The kids that I employ want to
do everything digital and scorn any sort of trimpot-like thing. It's a
good thing that I'm still the boss.

--

John Larkin Highland Technology, Inc

lunatic fringe electronics

[Spehro Pefhany]

On 9 Jan 2016 09:44:08 -0800, the renowned Winfield Hill

I always hate doing lifetime buys- it feels like betting against
success on the upside and throwing money away on the downside.


--sp


--
Best regards,
Spehro Pefhany
Amazon link for AoE 3rd Edition: http://tinyurl.com/ntrpwu8
Microchip link for 2015 Masters in Phoenix: http://tinyurl.com/l7g2k48

[John Miles, KE5FX]

On Saturday, January 9, 2016 at 9:59:48 AM UTC-8, John Larkin wrote:
> I'm designing a new oscillator now, and prefer to use a small
> surface-mount trimmer cap to center the frequency. One could tweak
> that with a screwdriver in 30 seconds. The kids that I employ want to
> do everything digital and scorn any sort of trimpot-like thing. It's a
> good thing that I'm still the boss.
>

I had that attitude until fairly recently. I used a comparator and a trimpot in an application that could have been served by a SOT-23 microcontroller or something similar, on the grounds that it was silly to use a dedicated processor to drive a status indicator LED. Real Men still do analog, right?

I soon realized that since nobody buys trimpots anymore, nobody bothers making them very well, at least not at a price I want to pay. The second lot of trimpots I bought (presumably legitimate Bourns 3223W-1-102E parts, purchased from Mouser) exhibited a 25% failure rate as soon as they were soldered to the board.

So your less-experienced engineers may be right on this one. It seems that trimpots and trimmer capacitors are now considered weird parts. I'm learning not to use weird parts when I don't absolutely have to, even if it makes me feel like less of a Real Man.

-- john, KE5FX

[John Larkin]

On Sat, 9 Jan 2016 14:33:53 -0800 (PST), "John Miles, KE5FX"
<jmi...@gmail.com> wrote:

>On Saturday, January 9, 2016 at 9:59:48 AM UTC-8, John Larkin wrote:
>> I'm designing a new oscillator now, and prefer to use a small
>> surface-mount trimmer cap to center the frequency. One could tweak
>> that with a screwdriver in 30 seconds. The kids that I employ want to
>> do everything digital and scorn any sort of trimpot-like thing. It's a
>> good thing that I'm still the boss.
>>
>
>I had that attitude until fairly recently. I used a comparator and a trimpot in an application that could have been served by a SOT-23 microcontroller or something similar, on the grounds that it was silly to use a dedicated processor to drive a status indicator LED. Real Men still do analog, right?

Some people still do analog. It's not as popular as it used to be. But
customers still need to connect to real-world signals. It's fun to
interview recent EE grads who barely know how a 2-resistor voltage
divider works, much less a transistor.

If you used the sot23 uP, how would you set the trip level?

>
>I soon realized that since nobody buys trimpots anymore, nobody bothers making them very well, at least not at a price I want to pay. The second lot of trimpots I bought (presumably legitimate Bourns 3223W-1-102E parts, purchased from Mouser) exhibited a 25% failure rate as soon as they were soldered to the board.
>
>So your less-experienced engineers may be right on this one. It seems that trimpots and trimmer capacitors are now considered weird parts. I'm learning not to use weird parts when I don't absolutely have to, even if it makes me feel like less of a Real Man.
>
>-- john, KE5FX

We've used tens of thousands of the Bourns 3314G pots with no
problems. They don't need to be programmed, don't need JTAG
interfaces, none of that stuff. If you keep the required resolution at
a half per cent or so, they are quick and easy to set. Around 0.1%,
people can spend time tweaking.

I need to coarse-tune an LC oscillator, to get things into the pull
range of a varicap. A small trim cap is ideal. I've used the Maxim
digital capacitor in the past, but it's EOL now. It needed a couple of
pages of code, too.

[Winfield Hill]

John Larkin wrote...
>
> Some people still do analog. ...

I'm still a huge analog fan. I've adopted a new
mantra, the world doesn't run on 3.3 volts. But ...
if you have a controller in your design, and if
anything in its program needs to be touched during
initial test & setup, then why not have calibration
and tweaking functions under digital control?


--
Thanks,
- Win

[John Larkin]

On 9 Jan 2016 18:58:20 -0800, Winfield Hill <hi...@rowland.harvard.edu>
wrote:

Most of our products have an ARM, and a cal table in flash, and
usually an FPGA or three. The issue here is a VCO that needs to have
low phase noise and be very temperature stable. Varicaps have horrible
tempcos that vary with the capacitance, so I want to keep the pull
range small. So I want to tweak the gross resonant frequency to make
up for the L and C tolerances. The options seem to be either a 32-step
digital capacitor (like the EOL Maxim part, which I've used before) or
a tiny ceramic trimmer cap.

Turning the cap with a screwdriver once, at the start of the
production test, doesn't seem like a terrible burden. Some of the
objections against trimpots and trim caps simply aren't rational.

We use one tiny trimpot in gain-setting applications that works fine
past 1 GHz. It costs 11 cents.

[John Miles, KE5FX]

On Saturday, January 9, 2016 at 4:50:13 PM UTC-8, John Larkin wrote:
> If you used the sot23 uP, how would you set the trip level?

The comparator was used to turn an RGB LED from blue to white to indicate when an OCXO reached operating temperature. Instead of tweaking a trimpot during production to set the threshold, it would probably have been reasonable for a controller to watch for the current to fall to X% of its cold-start value. I could have programmed it in-system as part of the mastering process for the FPGA and USB firmware. For that matter, the FPGA itself could have done the job with a hacked single-slope ADC.

Ultimately even the trimpot was overkill because the OCXOs behaved consistently enough to work with a fixed pair of resistors. But the original plan was to support a couple of OCXOs from different vendors, so the trimpot made more sense initially.

-- john, KE5FX

[Robert Baer]

Do "both" in one step: Do the layout with a number of capacitors in
parallel and laser trim(cut) links to remove appropriate capacitor, OR
if amount of capacitance change is rather small, just trim away some of
the copper pad of the capacitor.
Digital trimming .. their way,and your way at the same time. No value
drift.

[Phil Hobbs]

Well, you'd have to have different BOMs for the two versions, so you
could just put in the right resistors for each OCXO.

[Jim Thompson]

On 9 Jan 2016 18:58:20 -0800, Winfield Hill <hi...@rowland.harvard.edu>
wrote:

Yep. More and more of my "Analog" chips have built-in
self-calibrating, controlled, most recently, by a uP via an SPI bus.
So there's a significant digital content on the chip... mostly "quiet"
except during a calibration cycle.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

[John Larkin]

You could just sense the current and turn on the LED below some value.
There's probably a value that works with all your OCXOs.

I like an LED that lights up proportional to oven current, bright at
turn-on and getting dim after warmup. More drama.

I guess a uP could have an algorithm that observes the current and
decides when the oven has stabilized. But that has lots of corner
cases. Seems like a lot of work for one LED.

It could easily cost us kilobucks to stock a small uP, get the
development and programming tools, write and test the code, and put
all that into production. That extends time-to-market, too. That might
make sense if we planned to build tens of thousands of units.
Otherwise, stuff one resistor alongside the selected OCXO.

[John Larkin]

The laser trimmer would cost 10s of kilobucks, and needs to be
programmed, and the board would have to be powered up and communicated
with while it's in the laser trimmer. And a laser can cut things but
it can't un-cut them.

A small screwdriver seems like a better investment.

The Murata trim caps that we stock are 3x3 mm and cost 25 cents.

[John Miles, KE5FX]

On Sunday, January 10, 2016 at 5:53:05 AM UTC-8, Phil Hobbs wrote:
> Well, you'd have to have different BOMs for the two versions, so you
> could just put in the right resistors for each OCXO.

There were some variations within the same OCXO P/N, though. The (proportional) ovens were set at the factory to target each crystal's individual turnover temperature. The spread in post-warmup current consumption at 25C wasn't very large, but it was there, and it seemed like a good idea to accommodate it with a trimmer. The ability to work with different OCXOs was a secondary consideration.

-- john, KE5FX

[John Miles, KE5FX]

On Sunday, January 10, 2016 at 10:12:30 AM UTC-8, John Larkin wrote:
> I like an LED that lights up proportional to oven current, bright at
> turn-on and getting dim after warmup. More drama.

I thought about doing that but it seemed too vague to be useful. The idea was to be able to tell at a glance if the OCXO was finished with its initial warmup. A simple 2-minute time delay would have been fine, but nooooo....

> I guess a uP could have an algorithm that observes the current and
> decides when the oven has stabilized. But that has lots of corner
> cases. Seems like a lot of work for one LED.

There was also some overshoot when the oven controller reached its target temperature. Already having gone full Steve Jobs with my OCXO monitor LED, I didn't want it to waffle between blue and white for several seconds while it rang down. Naturally the extent and duration of the ringing varied from one part to the next. At the same time I didn't want to use a lot of hysteresis on the comparator in case the oven came back on during operation, so the trimmer setpoint was fairly critical. There were at least a half-dozen better ways to implement this "feature," but it just didn't seem to be worth much thought at the time.

This particular board already had a 48 MHz 8051 and a 1.2M-gate equivalent FPGA, so if I had it to do over again, I'd read the current-sense resistor with a cheap ADC.

-- john, KE5FX

[John Larkin]

Do you really need an OCXO heater status LED?

[John Miles, KE5FX]

On Sunday, January 10, 2016 at 12:18:03 PM UTC-8, John Larkin wrote:
> Do you really need an OCXO heater status LED?

Hard to say, always room for debate there. Nobody really needs anything but air, food, water, and shelter. But life's little luxuries are what make our existence bearable, and I count OCXO heater status LEDs among them.

More seriously, the LED serves to discourage users from trying to measure phase noise and Allan deviation with a box that they just powered up 12 seconds earlier. And it's a useful sanity check for one of the more important components in the instrument.

I really like RGB LEDs. You can communicate a surprising amount of information with just one of them, even allowing for color-blind users.

-- john, KE5FX

[Spehro Pefhany]

My liquid CPU cooler pump has a throbbing red LED. Not flashing, it's
PWM'd or something to make it throb. I guess red is the new blue.

"Need" is a somewhat fluid concept in the developed world. Like the
fluids (presumably mostly genunine Chinese tapwater plus antifungals)
my gag gift LED-lit computer speakers squirt internally, in time with
the sound.

[Winfield Hill]

John Larkin wrote...

>
> The Murata trim caps that we stock are 3x3 mm and cost 25 cents.

What's the p/n.


--
Thanks,
- Win

[John Larkin]

On 11 Jan 2016 05:20:04 -0800, Winfield Hill

<hi...@rowland.harvard.edu> wrote:

>John Larkin wrote...
>>
>> The Murata trim caps that we stock are 3x3 mm and cost 25 cents.
>
> What's the p/n.

MFR1 DIGIKEY 490-2008-1-ND
MFR2 MURATA TZY2Z2R5A001R00

The low values are NP0 +-300 PPM, which is mediocre but still better
than a varicap. The digital caps are better, ballpark 30 PPM.