EL7900

the RC to

actual

re's

onous

and

s, and

edge rate

lag from

et-up

e making

t scrambled

ttent

the

don't

ondition

're

hift

ous about

used an

could

and

cause

tious

an

I didn't say that it wasn't. I was just pointing out that the mechanism that you'd put together to get there was sub-optimal.

There's nothing uncivil about saying that. It's not something you want hear, but it is something that you need to hear. A more diplomatic form of words wouldn't have made it clear to you that you'd under- performed, and you might later have made a fool of yourself by presenting this again as an admirable piece of circuit design. Leaving you in that state really would have been uncivil ...

Not really. I confine my observations to the subject under discussion, while John tends to change his ground whenever he has been caught out.

Really? But that's not the output you are feeding into your filters.

Where's the vitriol? Pointing out that you've used 22nF and 11nF capacitors when a more recherche version of the filter can let you get away with a single capacitor value is scarcely vitriolic.

Scarcely relevant here. You adduced the circuit as an arbitrary example of your expertise in circuit design. It's obvious what it does, and equally obvious that it isn't the most elegant way of doing it.

Since the OP didn't want the filters in the first place, this is totally irrelevant - you didn't post the specfication to which you designed the circuit when you posted the circuit in this thread, and pulling some imagined specification out of thin air isn't exactly going to make the circuit look any less clumsy.

The counter and decoder are doing pretty much what you need. The output stage is obviously very different. It's more apples plus oranges, and job of editing out the oranges isn't exactly heroic.

Why should I bother? I've got nothing to prove.

-- Bill Sloman, Nijmegen

on

It shouldn't.

The last time I did something like that, I used 4016's as the switches. The 4017 output impedance is higher when the output is connected to the positive rail than it is when output is connected to the negative rail, which does mess things up further.

-- Bill Sloman, Nijmegen

That sort of thing seems to be the only expertise you have left. You haven't said anything useful about electronics in years.

You just have to persist in being a manic-depressive Napoleonic runt, don't you? ...Jim Thompson

John, I don't understand what is wrong with what you posted. Am I missing something?

John

The statement stands. You have nothing useful to say about electronics. Just coarse ranting and claims about how you used to be "The Greatest."

the question is why would you mess about with all those parts when a single ~1$ avr/pic etc. would do it in one part including an oscillator and the option to do use pwm to make it even better....

-Lasse

RC to

actual

and

rate

from

making

scrambled

don't

condition

about

used an

could

you could possible make a better approximation of a sinewave than the

-Lasse

ing the RC to

the actual

there's

chronous

al

ng and

se edge rate

illy

he lag from

t set-up

are making

get scrambled

rmittent

If the

you don't

e condition

you're

t shift

nxious about

ave used an

end could

its.

ree and

puts

ck cause

tentious

two

lop

And

ncy

ole

ver an

n,

he

ing,

ve"

it

You don't "require" me to do anything. You wanted me to give you a schematic, but since the point is obvious enough without it, you are out of luck.

Not really. It would be obvious to a competent engineer, which is what you are claim to be.

Come off it. There's been stuff I thought I've know that wasn't so, and when that happens I admit it. It doesn't happen often. There have been several occasions where you have misinterpreted what I've posted, and incorrectly claimed that I was wrong, when I wasn't, and these occasions do seem to figure largely in your memory.

He posts nonsensical twaddle lifted from "The Register" and I point out that it's nonsense, so he proceeds to claim that I don't design stuff. It's boring and predictable, and it wouldn't happen if I didn't point out that he was posting twaddle, but the offences involved are different.

-- Bill Sloman, Nijmegen

ar

Don Lancaster's "magic sine waves" list a whole menagery of better approximations. The theory is fine. Getting them to work in practice is another story.

-- Bill Sloman, Nijmegen

their

now.

So far it looks like Bessel (all pass) and state variable can do it. Though not without some parts matching or TC to look after.

?-/

Only the non-inverting low-pass S-K has exactly X1 _DC_ gain. Everything else I can think of requires component matching. ...Jim Thompson

Bessel is a filter response, not a topology. Most of our S-K filters are Bessel or Transitional Bessel, which has the Bessel time domain response but better stopband rolloff.

Bessel response is not "all pass."

State Variable filters will have a DC gain that depends on at least two resistors per 2nd order section. So an 8th order Bessel filter might have the gain depend of the values of 16 resistors. A second-order SV stage will typically have three or four opamps and 10 or 12 passives. A 2nd order S-K section is one opamp and four passives and a DC gain of 1.000 that doesn't depend on the resistor values.

There is a filter form that looks like

in------+------R-----+--------------out | | | | | C | | | | +----------MAGIC BOX | | | gnd

but I don't recall what it is.

Yep, reminds me, there are gyrator forms with X1 DC gain in spite of having two OpAmps. ...Jim Thompson

LTC made a switched-capacitor filter like that back in the early 90s. It worked fine (for S-C) but you couldn't clock-tune it very far because the RC pole stayed still.

Cheers

Phil Hobbs

Sure, but then the resistors affect the DC gain.

The NuHertz software does filters most people have never heard of. The guy who does this is brilliant. And crabby.

aps

...

Is he generally crabby, or is it just that you who get up his nose?

-- Bill Sloman, Nijmegen