Sallen-Key input Z

Sometimes when using Zverev, you wonder why he has so few significant figures in the tables. But then thinking of the year it was published, it makes sense.

The book also talks about how to make physical (passive) LCR filters.

Of course you can make filters with low-spec components, but the whole point of the discussion is that if you want the resulting circuit to have accurate DC gain, unity-gain Sallen-Key is a win.

Can you design me a four-opamp biquad or a leapfrog with 1-ppm/K DC gain stability using normal 1%, 200 ppm/K resistors?

It seems that hating S-K is one of those sacred cows like beta-dependence or goto statements.

Cheers

Phil Hobbs

No. You don't get nearly as much intuition out of using a program to do the work. Old fashioned techniques focus prioritize what's important and that is a good thing.

Jon

If you actually advanced an argument that was relevant for instrument design, I'd be all ears. I'm glad to learn from anybody.

Cheers

Phil Hobbs

Since you usually snip the stuff you're responding to, we have no idea who you're talking to, or what you're talking about.

--

John Larkin Highland Technology Inc

Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators

I'd be interested to know its name. I did find this:

$10k was real money back then. Would get you maybe three pretty decent new cars. A production worker made a little more than half that much in a whole year.

Jon

quoted text -

I don't think that's going to happen in my lifetime. We're pushing

"Then stop giving us reels of pepper!" ;-)

with

Hide quoted text -

We have binders of surface mount parts in various sizes that the vendors give us. There are about 100 of each value and perhaps 100 values in a binder. There are a few cut tape strips in small pouches attached to pages in each binder. It really is handy to have these available when modifying boards. I generally have a success rate of about 5 0402s (larger components are higher since it's harder for them to hide) placed properly before one takes a walkabout.

all

with

Hide quoted text -

This, too, shall pass.

Never had a failure that could be attributed to a wandering 0402.

That filter's noise peaks at 40 *microvolts* per root Hz, about a thousand times worse than a decent 6-pole S-K. And you were whining about noise peaking!

Like all switchcap filters, it will kick clock spikes everywhere and alias anything it can get, including anything on the power rails. Nobody who does serious instrumentation would use that piece-o-junk.

It doesn't look very DC accurate to me anyhow. The buffer, the suggested output, has up to 20 mV of offset. And I bet the "out" pin loads the series resistor by drifty amounts. Looks like it has latchup issues, too.

Try again.

We were just noting that a new Tek 545 (huge 30 MHz scope) with a plugin would buy a Chevy Impala. Now a Rigol or Saleg color digital scope costs about 1% of a decent car.

I was making cheap high-volume stuff... as in telephone filters. (Keep in mind... very early '70's). ...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

You aren't listening very carefully, or perhaps you just aren't taking seriously what other folks care about who aren't in your end of the business.

It's not true in general that you can do it with no additional components, at least not without compromising the DC gain accuracy. And since an instrument designer has lots more flexibility in some ways than a chip designer, sticking with unity-gain SK filters is a win in a great many instrument applications.

And since you don't seem to be very consistent about your noise objections--advocating the LT1062, forsooth, at a SNR cost of probably

In my world, adding more parts costs money. Particularly ultraprecise resistors.

Cheers

Phil Hobbs

Phil, You lost the thread. The OP was observing peculiarities due to Zin of S-K filters. Larkin, as usual turned it into something else. The original thread was NOT ABOUT DC gain... the "whole point of the discussion is that if you want the resulting circuit to have accurate DC gain" was NOT the topic... and like who gives a f***?

There's a high volume market for that? I don't think so. Stop pontificating... you have not the credentials.

Naaaah. I've "hated" S-K for 1/2 a Century... right after I trounced Larry Huelsman's "desensitized" method right in his face in front of a crowd of engineers at Motorola (no problem when you can do the math). My boss at the time, Jan Narud, thought it so funny I thought he was going to be sick, he was laughing so hysterically.

S-K is for crap analog filters where tolerances and unexpected peaking don't matter.

Get back on topic. ...Jim Thompson

Indeed. Give it a rest before proven Larkenesque. ...Jim Thompson

Yep. I have a guy that I don't share >:-} ...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

Yep. It took some arm twisting to get Dickson (Don Dickson, he was still alive then) to pony up, but I had created a major market with Teltone. ...Jim Thompson

nodes

You sure treasure your memories of trashing other people.

Moron. It was miso who bent the thread by trashing S-K filters.

And what's wrong with discussing SK filters, in a thread about SK filters?

So, design a 6 or 8-pole lowpass filter, non SK, that has single-digit PPM DC gain stability and single-digit microvolts offset stability.

Bet you can't.

nodes

Back on topic? Wot 'opes. This is SED, remember? Besides, name calling isn't an argument.

Cheers

Phil Hobbs