From: jl@glen--canyon.com
On Sat, 31 Jan 2026 17:33:42 +0100, Gerhard Hoffmann
wrote:
>Am 31.01.26 um 16:41 schrieb john larkin:
>> On Sat, 31 Jan 2026 16:25:29 +1100, Bill Sloman
>> wrote:
>>
>>> On 31/01/2026 8:53 am, john larkin wrote:
>>>> On Sat, 31 Jan 2026 03:51:16 +1100, Bill Sloman
>>>> wrote:
>>>>
>>>>> On 31/01/2026 3:04 am, john larkin wrote:
>>>>>> On Sat, 31 Jan 2026 02:18:25 +1100, Bill Sloman
>>>>>> wrote:
>>>>>>
>>>>>>> On 31/01/2026 12:43 am, Liz Tuddenham wrote:
>>>>>>>> Bill Sloman wrote:
>>>>>>>>
>>>>>>>>> On 30/01/2026 9:15 pm, Liz Tuddenham wrote:
>>>>>>>>>> Bill Sloman wrote:
>>>>>>>>>>
>>>>>>>>>> [...]
>>>>>>>>>>> The only electronics I did as a kid was to build a completely
passive
>>>>>>>>>>> crystal set
>>>>>>>>>> [...]
>>>>>>>>>>
>>>>>>>>>> I think we may quote that in replies to some of your future posts.
>>>>>>>>>
>>>>>>>>> It didn't include any parts with gain, or any power source. What's
your
>>>>>>>>> preferred description of the classic crystal set?
>>>>>>>>
>>>>>>>> The part that caught my eye was: " The only electronics I did as a
kid".
>>>>>>>> Many of us spent our childhood teaching ourselves electronics - so we
>>>>>>>> may remind you of this difference next time you start making
disparaging
>>>>>>>> remarks about other engineers' knowledge and abilities.
>>>>>>>
>>>>>>> John Larkin seems to think it gives you some kind of advantage.
>>>>>>
>>>>>> Of course it does. As there is a huge advantage to learning chess or
>>>>>> math or languages or soccer when you are young. Actually doing stuff
>>>>>> involves practical feedbacks and acquired instincts.
>>>>>
>>>>> Instincts are what you were born with. What you get from doing stuff is
>>>>> habits.
>>>>>
>>>>> Learning stuff too early can instill bad habits, and they are hard to
>>>>> unlearn.
>>>>>
>>>>> Languages aren't learned any faster if you learn them young, and some
>>>>> aspects of language can't be learned at all by very young kids.
>>>>>
>>>>>> University education seldom installs much in the way of instincts
>>>>>> either. It's too rigid and formalized, and too late.
>>>>>
>>>>> Since instincts are what you get with your genome, universities can't
>>>>> install them at all.
>>>>>
>>>>> Formal instruction at university is formal. It's mostly accompanied by
>>>>> practical classes, which are a lot less rigid.
>>>>>
>>>>> The complicated stuff that most people learn at university mostly can't
>>>>> be instilled into adolescents - some rare kids can learn it early, but
>>>>> they tend to be exceptionally clever and need exceptional power of
>>>>> concentration. About 30% of the undergraduate intake doesn't ever get
>>>>> any kid of degree, and probably shouldn't have started at all.
>>>>>
>>>>>>> If you taught yourself when you were a kid, you didn't have a
>>>>>>> well-qualified teacher.
>>>>>>
>>>>>> A mentor with instincts is great if you are lucky enough to have one.
>>>>>
>>>>> Instincts come from the genome. What good mentors have is experience,
>>>>> and some understanding of what that experience has taught them.
>>>>>
>>>>> Electronics has advanced a lot over the past fifty years, and mentors
>>>>> are correspondingly less useful as teachers.
>>>>>>> At least when I got into it, I did have a
>>>>>>> university library and book-shop to draw on and did get some advice
from
>>>>>>> people who really knew what they were doing.
>>>>>>
>>>>>> Obviously too late.
>>>>>
>>>>> What's obvious to you is what you want to see. Trump is even more deeply
>>>>> into wishful thinking than you are.
>>>>>
>>>>>>> I learned a lot when I started doing electronic engineering as my main
>>>>>>> job, and had some really skilled teachers and examplars, as a well as
>>>>>>> lot of colleagues who merely thought that they knew what they were
>>>>>>> doing, and earned a few disparaging remarks. A few disparaging remarks
>>>>>>> got published as comments in the Review of Scientific Instruments.
>>>>>>
>>>>>> I sometimes read RSI when it's available. The circuits are hilarious.
>>>>>
>>>>> They tend to be functional, rather than elegant, and not always all that
>>>>> up-to-date. I once got very rude about a paper lauding the use of 1Ok
>>>>> ECL which got published after ECLinPs had been around for a few years.
>>>>>
>>>>> 10k ECL was about four times faster than TTL/CMOS, but ECinPS was four
>>>>> times faster again. The same paper described a ripple carry counter
>>>>> where the carry propagation wasn't fast enough to match the maximum
>>>>> count rate claimed. No mention at all of a synchronous counter.
>>>>>
>>>>> It was a particularly horrible example, quite the worst I've ever seen.
>>>>
>>>> A true ripple counter is as fast as its first flop.
>>>
>>> Rubbish. The state of the outputs of a multistage ripple counter isn't
>>> useful until the increment has rippled through every stage.
>>
>>
>> Consider a frequency divider.
>>
>> Your prime motivation is to contradict, not to think. That's very
>> common.
>
>Are you soliloquizing?
>
>> A true ripple counter is as fast as its first flop.
>
>You know the count of a ripple counter when everything has come
>to a halt, including the carry chain.
A ripple counter has no carry chain. That's its charm. Of course
ripple counters rarely make sense in an FPGA. We do use sort-of ripple
counters in the PLLs.
>That is slower than the
>first flipflop.
>Even decoded outputs will feature transient wrong results.
That's actually interesting. I think if you decode count N with an AND
gate, here are no decode glitches below N.
>
>Look ahead carry has been invented, just for this.
>An LFSR is hard to beat but will very probably also need some
>combinatorial delays.
That's a tough way to make a counter.
>
>Gerhard
John Larkin
Highland Tech Glen Canyon Design Center
Lunatic Fringe Electronics
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
