Math intensive, transforms required, not for those who can't already do a hand-math filter design with Laplace (and need to fall back on FilterPro :-)
My favorite section is the "General Parameter Filters"... put in your pass-band/stop-band requirements in Z-transform notation (not the same as today's sampled data Z) and get out the result in Laplace.
I had an old TI calculator (~1977-78, TI-59?,
formatting link
with a docking station and a magnetic card reader, where I programmed splitting that Z-transform notation with output as quadratic chunks suitable for building in my p+1 version of state-variable filters. ...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 http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Though I rarely use S-K filters because their component sensitivities become gross at any Q>1, you _can_ easily calculate (BY HAND) the effect of OpAmp GBW on performance... try it sometime, the gain of an OpAmp is simply...
GBW/s
(You can ignore the DC "flattening" for purposes of analyzing AC performance.) ...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 http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Actually, the circuit is a joke. What John Fields ought to be synthesising - and what I originally thought that he was synthesising, until I looked at the waveforms in detail - is a "modified sine wave"
-- --
- - - - - -- --
where the nett signal being fed into U4, U5 and U6 is high for one third of the time, low for one third of the time and and zero for two intervening periods of one sixth of the time. This has no third harmonic content, and has been popular in cheap inverters for many years.
I don't know if there is an elegant way to synthesis this with Johnson counters - I can't think of one off the top of my head.
It's trivial with regular counters and bit of decoding logic.
--
Since the application is for a fixed frequency, there's no need for a
filter with a flat passband.
The 3rd harmonic is down more than 50 dB from the fundamental, so I
wouldn't expect a lot of trouble there, depending on the application,
of course.
The second harmonic, at ~ -30dB, might be more troublesome.
It's what you should have designed if you wanted to produce three close approximations to a sine wave precisely 120 degrees apart.
There are better digital approximations to a sine wave - Don Lancaster's magic sine waves list a whole lot of them - but the digital logic required to generate them gets bulky. Easy enough to realise in a cheap programmable logic device, but messy.
Having zero third harmonic content reduces the RC product of the filters you need to reduce the higher harmonic content of the digital first approximation to something "close enough" to a sine wave, without the tolerances on the Rs and Cs in the filters introducing more arbitrary phase differences than strictly necessary.
That you didn't have that in mind when you designed the circuit makes you close to incompetent
But it's more nearly sinusoidal than what you did produce, and correspondingly easy to filter.
Messy in LTSpice. This does show the divide-by-twelve that you'd need, even though it isn't doing quite the same thing. You'd have to put together a different collection of gates to generate a bunch of modified sine waves, but that wouldn't be all that difficult.
Slowman can't even fart without someone to assist him with parting his ass... two assistants, two flashlights, and at least one mirror ;-) ...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
formatting link
| 1962 | I love to cook with wine. Sometimes I even put it in the food.
If you'll fart, I'll light the match ;-) ...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 http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
The logic is a mess, A4 and A5 are superfluous, the phase3 filter input can be taken from Q-bar on A2 and all that's neede to ensure sensible startup and glitch recovery is an and gate taking A2 Q-bar and A3 Q and feeding it back to A1 CLR,
--
?? 100% natural
--- Posted via news://freenews.netfront.net/ - Complaints to news@netfront.net
Jim-out-of-touch-with-reality_Thompson has got me kill-filed, so he posted the above miss-estimation a little bit more than two hours after I'd posted a circuit satisfying John Fields' request. Admittedly, it was a circuit that I'd put together years earlier, so my "trivial" was based on more direct experience than such comments usually are, but it was perfectly adequate to illustrate that Jim's judgement is inadequate (as if that needed further illustration)..
Jim-out-of-touch-with-reality-Thompson hasn't quite taken on board that he's in Arizona and John Larkin is in California - the match has to be a little closer to the methane concentration to do anything useful.
Not good. The "Reset" input on the 4017 is an asynchronous clear, so this would produce a very narrow clear pulse - essentially the propagation delay from "reset" to Q7. There's no minimum propagation time. The typical at 5V is 265nsec, and the worst case minimum reset time is 260nsec, so you might manage to reset only some of the bistables.
And the resistor values don't look right either. You could simulate a
4017 in LTSpice - the data sheet includes a circuit digram in terms of stuff that's in the LTSpice library
Tor the correct ratio 12K should be 11.86K, but they're a bit low this means I'd have to consider the 4017 output impedance, but that seems to vary with supply voltage maybe just multiply by 1000 and hope.
--
?? 100% natural
--- Posted via news://freenews.netfront.net/ - Complaints to news@netfront.net
With three flip-flops and 6 legal states there are only 2 illegal states (101) and (010). If it starts in one of the illegal states the circuit will alternate between them on each clock pulse. The and gate traps one of them (101) and converts it to a legal state (001).
--
You need to go back and examine the output of my circuit a little more
closely, since then you may come to realize that's precisely what it
does produce.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.