Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:
> On 05/24/2017 09:13 PM, Steve Wilson wrote:
>> I reviewed my old archives. It turns out I used the HP 5326 in 1971.
>> This could reach 100ps using Time Interval Averaging, which was
>> excellent for the time.
>> HP released the 5370 around 1978. This could reach 20ps single shot
>> and 63 fs using 1e5 averaging. This took too long in production, so I
>> settled on 1e4 averaging. This gave 200 fs which was 500x better than
>> the original.
>> It is clear that by the early 1980's, many people were involved in
>> picosecond and femtosecond design and measurements.
> Yah. But not you, in 1970.
> Cheers
> Phil Hobbs
I never said 1970. I said 1971.
So I was 7 years off. Sue me.
My measurement was far more accurate than the HP 5326 could give.
And I still made millions. In 1980's dollars.
Had no clue how to handle it, and lost it all. But next time will be
different.
Reply by John Larkin●May 26, 20172017-05-26
On Fri, 26 May 2017 08:54:11 -0500, "Tim Williams"
<tiwill@seventransistorlabs.com> wrote:
>"John Larkin" <jjlarkin@highlandtechnology.com> wrote in message
>news:t7qdic9aaljjuvu9njqe8fusu5jj1ksv9e@4ax.com...
>> Nice book, "The Revenge of Analog." In one section, he talks about
>> super-high-tech companies that forbid electronic devices at design
>> meetings, and use whiteboards and paper.
>>
>
>In contrast, here's a possible vision of the future:
>https://youtu.be/F7yLL5fJxT4
>Works fine for art. I mean, engineering is art, too, so why not?
>
>I'd love to have waveforms at my fingertips.
I love oscilloscopes. I have about 60 of them.
>
>The ultimate in breadboarding: instant simulation results, without the
>trouble of soldering things together, and producing magic smoke.
Soldering together and measuring circuits is not only art, it's real.
I suspect that people who sell low millions of 15 cent parts will
never invest in supplying a really accurate electrical and thermal
model of their parts. They don't even edit out obvious blunders on the
data sheet, not years after being alerted to them.
>
>Downside: while we have the computational power to do it (if not necessarily
>the optimizations to avoid Amdahl's law*), SPICE development has been
>utterly stagnant over the past three decades. It would be foolish to think
>that would change "because VR".
>
>But one can hope.
>
>*Numerical simulations are timestepped, but there are some known shortcuts,
>like for quasi-periodic cycle-averaged systems, which are common enough to
>be worth addressing. Unfortunately they aren't in any free tools...
>
>Tim
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Reply by Jan Panteltje●May 26, 20172017-05-26
On a sunny day (Fri, 26 May 2017 08:54:11 -0500) it happened "Tim Williams"
<tiwill@seventransistorlabs.com> wrote in <og9brk$o7r$1@dont-email.me>:
>In contrast, here's a possible vision of the future:
>https://youtu.be/F7yLL5fJxT4
>Works fine for art. I mean, engineering is art, too, so why not?
>
>I'd love to have waveforms at my fingertips.
>
>The ultimate in breadboarding: instant simulation results, without the
>trouble of soldering things together, and producing magic smoke.
>
>Downside: while we have the computational power to do it (if not necessarily
>the optimizations to avoid Amdahl's law*), SPICE development has been
>utterly stagnant over the past three decades. It would be foolish to think
>that would change "because VR".
>
>But one can hope.
>
>*Numerical simulations are timestepped, but there are some known shortcuts,
>like for quasi-periodic cycle-averaged systems, which are common enough to
>be worth addressing. Unfortunately they aren't in any free tools...
Long time ago I tried walking into a building (where I was for the first time)
and hit the glass entrance door.
I was OK (made a loud bang, they were really concerned, told me I was number n who did that).
So, once we get used to 'virtual objects' we may get into the habit of ignoring those
and hit the real car or person ...
Then there could be no longer a difference between reality as we know it and what the machine makes us believe.
the Matrix, next you will grow up in a tube, get your brain interface,
wait, why need a brain at all, why need life.
SillyCon rules.
Maybe there is a planet somewhere where this transition did happen.
And an endless ever repeating simulation of life is all there is left.
...
Reply by Tim Williams●May 26, 20172017-05-26
"John Larkin" <jjlarkin@highlandtechnology.com> wrote in message
news:t7qdic9aaljjuvu9njqe8fusu5jj1ksv9e@4ax.com...
> Nice book, "The Revenge of Analog." In one section, he talks about
> super-high-tech companies that forbid electronic devices at design
> meetings, and use whiteboards and paper.
>
In contrast, here's a possible vision of the future:
https://youtu.be/F7yLL5fJxT4
Works fine for art. I mean, engineering is art, too, so why not?
I'd love to have waveforms at my fingertips.
The ultimate in breadboarding: instant simulation results, without the
trouble of soldering things together, and producing magic smoke.
Downside: while we have the computational power to do it (if not necessarily
the optimizations to avoid Amdahl's law*), SPICE development has been
utterly stagnant over the past three decades. It would be foolish to think
that would change "because VR".
But one can hope.
*Numerical simulations are timestepped, but there are some known shortcuts,
like for quasi-periodic cycle-averaged systems, which are common enough to
be worth addressing. Unfortunately they aren't in any free tools...
Tim
--
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Contract Design
Website: http://seventransistorlabs.com
Reply by John Larkin●May 25, 20172017-05-25
On 25 May 2017 05:49:02 -0700, Winfield Hill
<hill@rowland.harvard.edu> wrote:
>pcdhobbs@gmail.com wrote...
>>
>>> ah, you guys are fooling around with 50 to 300MHz parts,
>>> go play with Intersil's serious monolithic BJTs: HFA3096,
>>> HFA3127, HFA3128, HFA3046, HFA3101, HFA3102, etc. These
>>> have fT ratings up to 10GHz. HFA3134 and HFA3135, nice!
>>
>> I keep trying to use those for something cool, and somehow
>> they never quite fit. Disappointing Ree' and beta linearity,
>> mostly. Also V_A is only about 15V.
>
> OK, actually, I haven't found many useful applications
> either. Their operating current ratings are rather low.
The problem with a transistor with Ft=10 GHz is that beta=1 at 10 GHz!
I like phemts and mmics for fast stuff. Bipolars never seem to be as
fast as they promise, unless you tune them, and that's just RF.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Reply by John Larkin●May 25, 20172017-05-25
On 25 May 2017 04:38:45 -0700, Winfield Hill
<hill@rowland.harvard.edu> wrote:
>Steve Wilson wrote...
>>
>> No matter how painful or inconvenient, LTspice is still
>> orders of magnitude more useful than ASCII or hand-drawn
>> schematics.
>
> To the contrary, this thread has illustrated that
> LTspice drawings can be utterly useless. One can
> make a SPICE model do almost anything apparently,
> and reality has little to do with it. In contrast,
> a hand-drawn schematic, depending on who has drawn
> it, is more likely to have knowledgeable analytical
> analysis behind it, and therefore be believable and
> eminently useful.
Nice book, "The Revenge of Analog." In one section, he talks about
super-high-tech companies that forbid electronic devices at design
meetings, and use whiteboards and paper.
I just discovered some wonderful new pens, Uniball Vision 0.7 mm, in
colors. They increase my IQ about 5 points.
I design on D-size vellum, with Turquoise F pencils, and I'm pretty
productive.
There seems to be a hobby of sort of randomly throwing parts onto an
LT Spice schamatic and fiddling until something happens. I love LT
Spice, but I don't trust it. I've loved women like that.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Reply by Gerhard Hoffmann●May 25, 20172017-05-25
Am 25.05.2017 um 14:05 schrieb Winfield Hill:
> bitrex wrote...
>>
>> Ideally one could use one of those monolithic multi-transistors
>> on a single die, like the CA3096...it doesn't seem like you can
>> get 'em for any reasonable price anymore, though.
>
> Bah, you guys are fooling around with 50 to 300MHz parts,
> go play with Intersil's serious monolithic BJTs: HFA3096,
> HFA3127, HFA3128, HFA3046, HFA3101, HFA3102, etc. These
> have fT ratings up to 10GHz. HFA3134 and HFA3135, nice!
They were real fun in a time stretcher, that ft combined with
the fat space qualified ceramic flatpack with looong leads.
But they are about the only qualified RF PNPs that are left.
Gerhard
Reply by Winfield Hill●May 25, 20172017-05-25
pcdhobbs@gmail.com wrote...
>
>> ah, you guys are fooling around with 50 to 300MHz parts,
>> go play with Intersil's serious monolithic BJTs: HFA3096,
>> HFA3127, HFA3128, HFA3046, HFA3101, HFA3102, etc. These
>> have fT ratings up to 10GHz. HFA3134 and HFA3135, nice!
>
> I keep trying to use those for something cool, and somehow
> they never quite fit. Disappointing Ree' and beta linearity,
> mostly. Also V_A is only about 15V.
OK, actually, I haven't found many useful applications
either. Their operating current ratings are rather low.
--
Thanks,
- Win
Reply by ●May 25, 20172017-05-25
> Bah, you guys are fooling around with 50 to 300MHz parts,
> go play with Intersil's serious monolithic BJTs: HFA3096,
> HFA3127, HFA3128, HFA3046, HFA3101, HFA3102, etc. These
> have fT ratings up to 10GHz. HFA3134 and HFA3135, nice!
I keep trying to use those for something cool, and somehow they never quite fit.
Disappointing Ree' and beta linearity, mostly. Also V_A is only about 15V.
Now if I could get monolithic dual BFP640s, _that_ would be cool.
My latest gizmo uses a couple of those wrapped round each side of a MAT14 diff pair. It's a biased Darlington stage to return the base current to the collector to extend the effective f_T of the MAT14, and a common emitter stage looking at the MAT14's base and driving the Darlington's base to keep it still.
In the spherical-cow universe that extends the MAT14's f_T by about an order of magnitude while keeping its other nice properties.
We'll see how it works in a week or two.
Cheers
Phil
Reply by Winfield Hill●May 25, 20172017-05-25
bitrex wrote...
>
> Ideally one could use one of those monolithic multi-transistors
> on a single die, like the CA3096...it doesn't seem like you can
> get 'em for any reasonable price anymore, though.
Bah, you guys are fooling around with 50 to 300MHz parts,
go play with Intersil's serious monolithic BJTs: HFA3096,
HFA3127, HFA3128, HFA3046, HFA3101, HFA3102, etc. These
have fT ratings up to 10GHz. HFA3134 and HFA3135, nice!
--
Thanks,
- Win