Maker Pro
Maker Pro

Are active filters realistic at 30MHz?

J

Joel Kolstad

Jan 1, 1970
0
Just curious... has anyone used an active filter at 30MHz or higher? I'd be
looking for a second order bandpass filter up there, and using a state
variable topology I'm reading that the GBW of the amplifier needs to be at
least 3*Q*f0... and I can live with a Q of about 10-12.5, so that would
indicate a GBW in the ballpark of a GHz. Analog Devices will such you such
an op-amp, but is it going to very difficult to get the filter to perform
properly when the GBW is the same as the '3*Q*f' formula? Or does that
formula assume you'll spec the op-amp's GBW directly from it?

Other than LC filters, are there other common methods for HF range filters?

Thanks,
---Joel
 
J

Jim Thompson

Jan 1, 1970
0
Just curious... has anyone used an active filter at 30MHz or higher? I'd be
looking for a second order bandpass filter up there, and using a state
variable topology I'm reading that the GBW of the amplifier needs to be at
least 3*Q*f0... and I can live with a Q of about 10-12.5, so that would
indicate a GBW in the ballpark of a GHz. Analog Devices will such you such
an op-amp, but is it going to very difficult to get the filter to perform
properly when the GBW is the same as the '3*Q*f' formula? Or does that
formula assume you'll spec the op-amp's GBW directly from it?

Other than LC filters, are there other common methods for HF range filters?

Thanks,
---Joel

May I recommend you take a look at the gyrator-based filter structures
on my website? I would think a GHz GBW OpAmp will do just ducky at
30MHz.

...Jim Thompson
 
J

Joel Kolstad

Jan 1, 1970
0
Thanks Jim, will do. In fact I already came across your gyrator web page
while Googling -- you had quite the discussion with Win and Peter Brackett!
 
K

Ken Smith

Jan 1, 1970
0
Just curious... has anyone used an active filter at 30MHz or higher? I'd be
looking for a second order bandpass filter up there, and using a state
variable topology I'm reading that the GBW of the amplifier needs to be at
least 3*Q*f0... and I can live with a Q of about 10-12.5, so that would
indicate a GBW in the ballpark of a GHz. Analog Devices will such you such
an op-amp, but is it going to very difficult to get the filter to perform
properly when the GBW is the same as the '3*Q*f' formula? Or does that
formula assume you'll spec the op-amp's GBW directly from it?

Beware that the 3*Q*f0 thing means the GBP at the operating frequency.
Many op-amps fall a bit faster than a single pole. If you want the filter
to be very accurate, you may have to (gasp) trim it.

The f0 point tends to shift downwards in frequency as the op-amps add
excess phase to the system. You may be able to just compensate for the
typical values and let the variation in op-amps be a variation in the
performance. A little figuring and calculating is needed.
 
T

Tim Wescott

Jan 1, 1970
0
Joel said:
Just curious... has anyone used an active filter at 30MHz or higher? I'd be
looking for a second order bandpass filter up there, and using a state
variable topology I'm reading that the GBW of the amplifier needs to be at
least 3*Q*f0... and I can live with a Q of about 10-12.5, so that would
indicate a GBW in the ballpark of a GHz. Analog Devices will such you such
an op-amp, but is it going to very difficult to get the filter to perform
properly when the GBW is the same as the '3*Q*f' formula? Or does that
formula assume you'll spec the op-amp's GBW directly from it?

Other than LC filters, are there other common methods for HF range filters?
Crystal filters, if you need 0.01 to 0.1% relative bandwidths.

I suspect that an LC filter will be more frequency stable than anything
with an op-amp (sorry Jim), it'll be passive so you won't have to worry
about oscillations, and at 30MHz it won't be much bigger, if at all.

Whatcha trying to do?
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Joel Kolstad <JKolstad71HatesSpam@
Other than LC filters, are there other common methods for HF range
filters?

30 MHz is the border between HF and VHF.

Ceramic.

SAW (Surface Acoustic Wave).
 
J

Joerg

Jan 1, 1970
0
Hello Joel,

I would crack out Williams filter handbook and do it LC. Much cheaper,
and easier.

Regards, Joerg
 
J

Joel Kolstad

Jan 1, 1970
0
Tim Wescott said:
Whatcha trying to do?

Build an electronically tunabe notch filter for the entire HF band. :)

'Electronically tunable' can mean horrible things like motor driven L/C's,
but I'd really like to avoid that if possible.

---Joel
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Joel Kolstad <JKolstad71HatesSpam@
Yahoo.Com> wrote (in said:
Build an electronically tunabe notch filter for the entire HF band.

Heterodyne up to the 100 MHz band with a fixed oscillator, make a
tunable notch filter for 100 MHz to 127 MHz and heterodyne down again.
Or 500 MHz, if 27% bandwidth is too much.

Or heterodyne up with a tunable oscillator to a fixed 100 MHz notch
filter and heterodyne back down.
 
M

Michael Black

Jan 1, 1970
0
Tim said:
Crystal filters, if you need 0.01 to 0.1% relative bandwidths.

I suspect that an LC filter will be more frequency stable than anything
with an op-amp (sorry Jim), it'll be passive so you won't have to worry
about oscillations, and at 30MHz it won't be much bigger, if at all.

Whatcha trying to do?
That has to be the key question.

Whether or not gyrators are suitable up there, gyrators tend to be
used where the coil it replaces is way too bulky, and even where
the coil may cause stray pickup. These are audio issues, since a big
inductor (electrically and physically) can so easily pick up stray 60Hz,
and the large inductance can take up space.

At 30MHz, the coils become small, and they have too little inductance
to pick up that 60Hz. The active circuitry is likely to use up more
space than the inductors they replace. Expecially since a far more
complicated filter will be needed at 30MHz to get desired bandwidth.

One also wonders how the gyrators act under strong RF signals. Since
one reason for filters at RF is to ensure the active elements don't
see too much signal, does using active elements as inductors lower
the signal handling capability of the receiver?

When people ask about high frequency gyrators, they seem to generally
lack a grounding in RF, and are trying to extend audio concepts that
they know to a field they don't know.

If one needs a lot of selectivy, you use a crystal filter at the signal
frequency (if you can afford it), or go to a superheterodyne receiver
so the selectivity is at a frequency where it can be done easily, with
front end selectivity only there to limit what the active components
see.

One could always add some regeneration to front end filtering to
increase Q, but of course that doesn't help skirt selectivity.

Michael
 
J

Joel Kolstad

Jan 1, 1970
0
Hi John,

John Woodgate said:
Heterodyne up to the 100 MHz band with a fixed oscillator, make a
tunable notch filter for 100 MHz to 127 MHz and heterodyne down again.
Or 500 MHz, if 27% bandwidth is too much.

Yeah, I've thought of that, and it is quite attractive. The idea of an
active filter instead was to introduce less distortion in the whole process,
but I have no hard evidence for whether or not that's really the case
instead of just a gut feeling.

BTW, how were you planning to build your tunable 100-127MHz notch filter?
Varactor diodes in an LC filter?

---Joel
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Joel Kolstad <JKolstad71HatesSpam@
BTW, how were you planning to build your tunable 100-127MHz notch
filter? Varactor diodes in an LC filter?

I left that as an exercise for the student. If you have room, a trough
line or a cavity might be interesting. VERY sharp notch.
 
Top