On 25/03/2010 12:37 PM, dave.harper wrote:
On Mar 24, 3:56 pm, wrote:
On 24 Mar, 18:22, wrote:





On Mar 23, 3:00 pm, John Larkin

wrote:
On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper"

wrote:
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like.

A lowpass filter is appropriate. You're trying to kill the higher
(odd) harmonics but there's no signal below the fundamental, excepting
DC maybe.

A dual-stage RC lowpass would be pretty good, if you set the -3 dB
point around 2 KHz. A higher-order filter, active or LC, would be even
better. An LC filter isn't unreasonable at 2 KHz.

John

I modeled a few low pass filters, and think this is the best
solution. I have 3 RC low-pass filters and the output looks
relatively sine-like. Thanks to everyone that provided good
suggestions! Here's the schematic in LTSpice (comments appreciated):

An inductor really does make a big difference for these kind of jobs
and only costs pennies.
I've tacked on an LC filter to your drawing.
Output is 5Vpkpk sine at very low distortion.
(filter is a 3rd order, Chebyshev, low pass, with 6db passband ripple
and a Zero notch at 6kHz. Components rounded to preferred values)
[replacement for LTlist...]

That is a nicer output, but that also requires a 15mH coil. Wouldn't
that be a little big? This will be going into a small payload
container with 3" diameter.

Amateur rocket?

"Don Lancaster" wrote in message
...
Bob Eld wrote:
"dave.harper" wrote in message

...
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like. I've also considered using a counter feeding a bank of
resistors, but finding a method for it to start counting up with it
hits 0 and down when it hits the high value might be more difficult.

Originally, I had used a wein-bridge oscillator with op-amps to make
the sine wave, but due to the environment, EMI was a problem and it
damped out the oscillations (even with ferrite beads and modest
shielding). So I'd like to avoid using op-amps (since they are
apparently sensitive to EMI) if possible.

Any suggestions?

Thanks in advance!
Dave

I would make a five pole elliptic low pass active filter breaking at
about 3
kHz with a deep 80 dB notch at the third harmonic of 6 kHz. This is a
dual
op amp (2 section) solution and uses a small number of resistors and
caps.
You should be able to attenuate all of the harmonics at least 80 dB
leaving
a pretty good sine wave.

Parallax invented the ultimate in mind-numbingly simple sinewave
generators many years ago. SIX BYTES !!! of working code on a PIC!!!!

Start with page 85.3 of http://www.tinaja.com/glib/hackar4.pdf
--
Many thanks,

Don Lancaster voice phone: (928)428-4073
Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552
rss: http://www.tinaja.com/whtnu.xml email:

Please visit my GURU's LAIR web site at http://www.tinaja.com

Thanks Don. Truly weird! I plan to code it up and play around with this when
time permits. So, the original poster could put this into a single chip PIC
and forget about his square wave source. He'd still need to run a D toA to
get an analog output, would he not? Within that algorithm he could also pull
out a square wave too if needed. Frequency could be anything within the
speed of the processor with appropriate, even externally controlled internal
delays and of course be crystal controlled and very stable. Hmmm. I already
have a use for this. Thanks again.
Bob

On 25 Mar, 02:37, "dave.harper" wrote:
On Mar 24, 3:56*pm, wrote:





On 24 Mar, 18:22, "dave.harper" wrote:

On Mar 23, 3:00*pm, John Larkin

wrote:
On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper"

wrote:
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). *It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. *One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like.

A lowpass filter is appropriate. You're trying to kill the higher
(odd) harmonics but there's no signal below the fundamental, excepting
DC maybe.

A dual-stage RC lowpass would be pretty good, if you set the -3 dB
point around 2 KHz. A higher-order filter, active or LC, would be even
better. An LC filter isn't unreasonable at 2 KHz.

John

I modeled a few low pass filters, and think this is the best
solution. *I have 3 RC low-pass filters and the output looks
relatively sine-like. *Thanks to everyone that provided good
suggestions! *Here's the schematic in LTSpice (comments appreciated):

An inductor really does make a big difference for these kind of jobs
and only costs pennies.
I've tacked on an LC filter to your drawing.
Output is *5Vpkpk sine at very low distortion.
(filter is a 3rd order, Chebyshev, low pass, with 6db passband ripple
and a Zero notch at 6kHz. Components rounded to preferred values)
[replacement for LTlist...]

That is a nicer output, but that also requires a 15mH coil. *Wouldn't
that be a little big? *This will be going into a small payload
container with 3" diameter.

They're about the size of 0.4W resistors but if you're stuck for space
then it's not an option.
I had in mind this ... 'Epcos' part# B82144A2156J000
A pic of a similar item ...
http://uk.farnell.com/epcos/b78108s1...646?Ntt=608646
Or a different body style but more expensive ...
http://uk.farnell.com/murata-power-s...44?Ntt=1077044

On Mar 25, 10:57*am, wrote:
On 25 Mar, 02:37, "dave.harper" wrote:



On Mar 24, 3:56*pm, wrote:

On 24 Mar, 18:22, "dave.harper" wrote:

On Mar 23, 3:00*pm, John Larkin

wrote:
On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper"

wrote:
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). *It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. *One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like.

A lowpass filter is appropriate. You're trying to kill the higher
(odd) harmonics but there's no signal below the fundamental, excepting
DC maybe.

A dual-stage RC lowpass would be pretty good, if you set the -3 dB
point around 2 KHz. A higher-order filter, active or LC, would be even
better. An LC filter isn't unreasonable at 2 KHz.

John

I modeled a few low pass filters, and think this is the best
solution. *I have 3 RC low-pass filters and the output looks
relatively sine-like. *Thanks to everyone that provided good
suggestions! *Here's the schematic in LTSpice (comments appreciated):

An inductor really does make a big difference for these kind of jobs
and only costs pennies.
I've tacked on an LC filter to your drawing.
Output is *5Vpkpk sine at very low distortion.
(filter is a 3rd order, Chebyshev, low pass, with 6db passband ripple
and a Zero notch at 6kHz. Components rounded to preferred values)
[replacement for LTlist...]

That is a nicer output, but that also requires a 15mH coil. *Wouldn't
that be a little big? *This will be going into a small payload
container with 3" diameter.

They're about the size of 0.4W resistors but if you're stuck for space
then it's not an option.
I had in mind this ... 'Epcos' part# *B82144A2156J000
A pic of *a similar item ...http://uk.farnell.com/epcos/b78108s1...-4700uh/dp/608...
Or a different body style but more expensive ...http://uk.farnell.com/murata-power-s...ductor-15mh/dp...

Hm, what about this one? Looks promising:

http://search.digikey.com/scripts/Dk...me=811-2058-ND

On 25 Mar, 20:26, "dave.harper" wrote:
On Mar 25, 10:57*am, wrote:





On 25 Mar, 02:37, "dave.harper" wrote:

On Mar 24, 3:56*pm, wrote:

On 24 Mar, 18:22, "dave.harper" wrote:

On Mar 23, 3:00*pm, John Larkin

wrote:
On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper"

wrote:
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). *It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. *One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like.

A lowpass filter is appropriate. You're trying to kill the higher
(odd) harmonics but there's no signal below the fundamental, excepting
DC maybe.

A dual-stage RC lowpass would be pretty good, if you set the -3 dB
point around 2 KHz. A higher-order filter, active or LC, would be even
better. An LC filter isn't unreasonable at 2 KHz.

John

I modeled a few low pass filters, and think this is the best
solution. *I have 3 RC low-pass filters and the output looks
relatively sine-like. *Thanks to everyone that provided good
suggestions! *Here's the schematic in LTSpice (comments appreciated):

An inductor really does make a big difference for these kind of jobs
and only costs pennies.
I've tacked on an LC filter to your drawing.
Output is *5Vpkpk sine at very low distortion.
(filter is a 3rd order, Chebyshev, low pass, with 6db passband ripple
and a Zero notch at 6kHz. Components rounded to preferred values)
[replacement for LTlist...]

That is a nicer output, but that also requires a 15mH coil. *Wouldn't
that be a little big? *This will be going into a small payload
container with 3" diameter.

They're about the size of 0.4W resistors but if you're stuck for space
then it's not an option.
I had in mind this ... 'Epcos' part# *B82144A2156J000
A pic of *a similar item ...http://uk.farnell.com/epcos/b78108s1...-4700uh/dp/608...
Or a different body style but more expensive ...http://uk.farnell.com/murata-power-s...ductor-15mh/dp...

Hm, what about this one? *Looks promising:

http://search.digikey.com/scripts/Dk...tail&name=811-....

Yep. Fine!.
(DigiKey 811-1302-ND for the 15mH)

On Thu, 25 Mar 2010 13:26:27 -0700 (PDT), "dave.harper"
wrote:

On Mar 25, 10:57*am, wrote:
On 25 Mar, 02:37, "dave.harper" wrote:



On Mar 24, 3:56*pm, wrote:

On 24 Mar, 18:22, "dave.harper" wrote:

On Mar 23, 3:00*pm, John Larkin

wrote:
On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper"

wrote:
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). *It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. *One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like.

A lowpass filter is appropriate. You're trying to kill the higher
(odd) harmonics but there's no signal below the fundamental, excepting
DC maybe.

A dual-stage RC lowpass would be pretty good, if you set the -3 dB
point around 2 KHz. A higher-order filter, active or LC, would be even
better. An LC filter isn't unreasonable at 2 KHz.

John

I modeled a few low pass filters, and think this is the best
solution. *I have 3 RC low-pass filters and the output looks
relatively sine-like. *Thanks to everyone that provided good
suggestions! *Here's the schematic in LTSpice (comments appreciated):

An inductor really does make a big difference for these kind of jobs
and only costs pennies.
I've tacked on an LC filter to your drawing.
Output is *5Vpkpk sine at very low distortion.
(filter is a 3rd order, Chebyshev, low pass, with 6db passband ripple
and a Zero notch at 6kHz. Components rounded to preferred values)
[replacement for LTlist...]

That is a nicer output, but that also requires a 15mH coil. *Wouldn't
that be a little big? *This will be going into a small payload
container with 3" diameter.

They're about the size of 0.4W resistors but if you're stuck for space
then it's not an option.
I had in mind this ... 'Epcos' part# *B82144A2156J000
A pic of *a similar item ...http://uk.farnell.com/epcos/b78108s1...-4700uh/dp/608...
Or a different body style but more expensive ...http://uk.farnell.com/murata-power-s...ductor-15mh/dp...

Hm, what about this one? Looks promising:

http://search.digikey.com/scripts/Dk...me=811-2058-ND

---
How about this:

Version 4
SHEET 1 880 680
WIRE 208 0 16 0
WIRE 208 16 208 0
WIRE 16 32 16 0
WIRE 16 128 16 112
WIRE 208 144 208 96
WIRE 336 144 208 144
WIRE 208 160 208 144
WIRE 16 256 16 208
WIRE 208 256 208 224
WIRE 208 256 16 256
WIRE 16 352 16 256
FLAG 16 352 0
SYMBOL ind 192 0 R0
SYMATTR InstName L1
SYMATTR Value .00932
SYMATTR SpiceLine Rser=110
SYMBOL cap 192 160 R0
WINDOW 0 42 38 Left 0
WINDOW 3 26 65 Left 0
SYMATTR InstName C1
SYMATTR Value .68e-6
SYMBOL voltage 16 112 R0
WINDOW 3 24 104 Invisible 0
WINDOW 39 0 0 Left 0
SYMATTR Value PULSE(0 5 0 1e-7 1e-7 250e-6 500e-6)
SYMATTR InstName V1
SYMATTR Value2 AC 1
SYMBOL res 0 16 R0
SYMATTR InstName R1
SYMATTR Value 10
TEXT 40 288 Left 0 !.tran .1
TEXT 40 320 Left 0 !;ac oct 256 10 10000



The coil is a Caddell-Burns 6740-61 (6.8 - 12mH) variable inductor:

http://www.caddell-burns.com/PDF/A1042.pdf

You can use it to tune to 2kHz, dead nuts, but that tunability
comes at a price:

http://www.caddell-burns.com/priceLi..._List_Pg_2.pdf

and the cap is any garden variety 0.68µF polyester cap.

JF

Try a high value resistor, say 47k into a pair of reverse paralleled germanium diodes to earth. The square law "knee" will clip the signal into a psuedo sine, around 1 volt p to p. Simple, and it works at these low frequencies. Three components, give it a try. cheers

On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper" wrote:

I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like. I've also considered using a counter feeding a bank of
resistors, but finding a method for it to start counting up with it
hits 0 and down when it hits the high value might be more difficult.

Originally, I had used a wein-bridge oscillator with op-amps to make
the sine wave, but due to the environment, EMI was a problem and it
damped out the oscillations (even with ferrite beads and modest
shielding). So I'd like to avoid using op-amps (since they are
apparently sensitive to EMI) if possible.

Any suggestions?

Thanks in advance!
Dave

With a 10% of fc range i would use an integrator first then use an
active RC bandpass filter.

On Wed, 24 Mar 2010 14:02:12 -0700, Tim Wescott wrote:

dave.harper wrote:
On Mar 23, 3:00 pm, John Larkin
wrote:
On Tue, 23 Mar 2010 12:39:59 -0700 (PDT), "dave.harper"

wrote:
I'm trying to come up with a filter or converter that takes a square
wave input and outputs something closer to a sign wave (varying
between ~1.9kHz and 2.1kHz). It can have some distortion, but I'm
trying to eliminate the sharp leading and trailing edge. One option
I'm pursuing is a bandpass filter (2 caps and 2 resistors), which
looks to give a reasonable output, but still not quite as smooth as
I'd like.
A lowpass filter is appropriate. You're trying to kill the higher
(odd) harmonics but there's no signal below the fundamental, excepting
DC maybe.

A dual-stage RC lowpass would be pretty good, if you set the -3 dB
point around 2 KHz. A higher-order filter, active or LC, would be even
better. An LC filter isn't unreasonable at 2 KHz.

John

I modeled a few low pass filters, and think this is the best
solution. I have 3 RC low-pass filters and the output looks
relatively sine-like. Thanks to everyone that provided good
suggestions! Here's the schematic in LTSpice (comments appreciated):

If a well-designed Wien bridge oscillator gets hammered by RFI, you're
somewhat depending on luck that an active filter doesn't have the same
problem.

How many _well designed_ Wien bridge oscillators have you seen that get
hammered by EMI/RFI?