These rather interesting papers and thesis show a design of
varactor-tuned RF bandpass filter that spans a wider range than any
previous one. I hope folk here will find it interesting.

30 MHz constant bandwidth in the frequency range from 380 MHz to 920 MHz
- nice!

"The filter is thermally stable. It has 1 dB compression point of 10 dBm
and third-order intercept point of 21 dBm."

The IP2 figure is relatively low, probably because only single diode
used, not two in anti-phase. The mixing products should be outside the
pass-band, but still...

I've included the direct link to the PDFs on sci-hub:

<https://zero.sci-hub.se/4394/15d8435eae1ada4e02c929c802186ebc/golaszewski2015.pdf?download=true>

<https://zero.sci-hub.se/4394/7646ae21e7db2554018653da972e4837/golaszewski2015.pdf?download=true>

Same author's thesis (in Polish - Google translate works though):
<https://repo.pw.edu.pl/docstore/download/WUTd9e29764e9a845a398918cc67a26bc66/Arkadiusz+Go%C5%82aszewski+-+Praca+magisterska.pdf>

The varactors he used:
<https://www.skyworksinc.com/-/media/SkyWorks/Documents/Products/1601-1700/201953A.pdf>/

Clifford Heath

Clifford Heath wrote:
==================
> These rather interesting papers and thesis show a design of
> varactor-tuned RF bandpass filter that spans a wider range than any
> previous one. I hope folk here will find it interesting.
>
> 30 MHz constant bandwidth in the frequency range from 380 MHz to 920 MHz
> - nice!
>
> "The filter is thermally stable.

** Errr - not really.

Average 70kHz drift per degree C.

....... Phil

On 9/12/21 3:55 pm, Phil Allison wrote:
> Clifford Heath wrote:
> ==================
>> These rather interesting papers and thesis show a design of
>> varactor-tuned RF bandpass filter that spans a wider range than any
>> previous one. I hope folk here will find it interesting.
>>
>> 30 MHz constant bandwidth in the frequency range from 380 MHz to 920 MHz
>> - nice!
>>
>> "The filter is thermally stable.
>
> ** Errr - not really.
>
> Average 70kHz drift per degree C.

That's not too bad at upwards of 400MHz, it's under 0.02%/degree?
Also pretty easy to calibrate out using a lookup table too.

Clifford Heath wrote:
=====================
>
> >> These rather interesting papers and thesis show a design of
> >> varactor-tuned RF bandpass filter that spans a wider range than any
> >> previous one. I hope folk here will find it interesting.
> >>
> >> 30 MHz constant bandwidth in the frequency range from 380 MHz to 920 MHz
> >> - nice!
> >>
> >> "The filter is thermally stable.
> >
> > ** Errr - not really.
> >
> > Average 70kHz drift per degree C.

> That's not too bad at upwards of 400MHz, it's under 0.02%/degree?

** Yawnnnnnnn.....

> Also pretty easy to calibrate out using a lookup table too.

** Thing is, its ONLY an IF filter.

A bit of temp drift hardly matters, long as the BW is consistent.

The same drift in local oscillator would be fatal, in many apps.

FYI

I used to build broadcast FM tuners with varicaps.
Drift was a serious issue.

...... Phil

On 9/12/21 5:56 pm, Phil Allison wrote:
> Clifford Heath wrote:
> =====================
>>
>>>> These rather interesting papers and thesis show a design of
>>>> varactor-tuned RF bandpass filter that spans a wider range than any
>>>> previous one. I hope folk here will find it interesting.
>>>>
>>>> 30 MHz constant bandwidth in the frequency range from 380 MHz to 920 MHz
>>>> - nice!
>>>>
>>>> "The filter is thermally stable.
>>>
>>> ** Errr - not really.
>>>
>>> Average 70kHz drift per degree C.
>> That's not too bad at upwards of 400MHz, it's under 0.02%/degree?
> ** Yawnnnnnnn.....

Go away and break some more toasters Phil. We're talking MHz, not KHz
here. If you lack the imagination to see what use such a tunable filter
might be, you should just keep quiet.

There are many general-coverage receivers and handheld transceivers that
have bugger-all input filtering, so their receive response is full of
images due to LO harmonics interacting with mixer non-linearities and
intermodulation in the front-end amplifiers. No amount of digital
cleverness can get rid of these false signals once they're in the signal
bandwidth.

This kind of filter is almost trivially simple but would dramatically
improve the performance of these radios.

I'm surprised Gerhard hasn't responded yet...?

Clifford Heath

Cue dummy-spit from PA. Don't bother PA, nobody cares, however angry you
get.

Clifford Heath = Over Snipping Arsehole:
--------------------------------------------------------
>
>>>> "The filter is thermally stable.
> >>>
> >>> ** Errr - not really.
> >>>
> >>> Average 70kHz drift per degree C.
> >> That's not too bad at upwards of 400MHz, it's under 0.02%/degree?

> > ** Yawnnnnnnn.....

> Go away and break some more toasters Phil.

** Go fuck you mother.

> We're talking MHz, not KHz

** Why did you snip my post ?????

Purely so you could post a load of insane abuse ?

What a nasty, fucking prick you are.

...... Phil

On Wednesday, December 8, 2021 at 8:35:27 PM UTC-8, Clifford Heath wrote:
> I've included the direct link to the PDFs on sci-hub:

Thanks for the pointer.. Neat stuff, considering how expensive tunable
filter ICs are.

I have a sack of varicaps and an application in mind, so I might have to
give this topology a shot. Phase stability near midband is critical, but
I don't care about behavior near the passband edges.

-- john, KE5FX

On 10/12/21 12:59 pm, John Miles, KE5FX wrote:
> On Wednesday, December 8, 2021 at 8:35:27 PM UTC-8, Clifford Heath wrote:
>> I've included the direct link to the PDFs on sci-hub:
>
> Thanks for the pointer.. Neat stuff, considering how expensive tunable
> filter ICs are.
>
> I have a sack of varicaps and an application in mind, so I might have to
> give this topology a shot. Phase stability near midband is critical, but
> I don't care about behavior near the passband edges.

It seems to be related to the capacitance variation with frequency
that's used in interdigitated filters and capacitors, etc. This means
that the capacitance variation changes the coupling between the two
central inductors. A friend is currently trying to figure out how to
simulate these, since the effects are basically immune to analysis.