RF switches

[garyr]

I'm trying to build a 80/40 meter receiver and I'd like to be able to
switch the RF input between the two bandpass filters. This sort of thing is
usually done with relays but I'd like to avoid that if possible. A
SN74CBT3306 would be ideal but I think the on-resistance would be too high,
thought it's hard to tell from the data sheet. Are there other devices that
would be suitable for this application? What do you think about using
2N7000's for this job?

The input and output capacitors of the double-tuned filters are 2200 & 2700
pF so the stray capacitance of the switching circuitry would not affect the
response very much.

Thanks in advance

[Michael Black]

I don't recall every seeing CMOS switches used for switching LC cicuits.
Maybe for switching stages, but I can't think of an example of that.

If it's not a switch, which has the disadvantage of not being close to
where the circuit needs switching, it's a relay. If it's not a relay,
then switching diodes are called in. Lots of people were using small
signal diodes for switching in the early days, then PIN diodes came along
they got the glory. The one issue seems that under some circumstances,
they can add distortion.

In the very old days, the tuned circuits were plugged in, very troublesome
but it avoided switches. Once semicondcutors came along, they were cheap
and small enough that duplication often made sense, switch whole stages,
including the tuned circuits, rather than switching the coils alone. That
gives you more leeway, with active stages in there, any loss from CMOS
switches goes way down.

Michael VE2BVW

[Arid ace]

In receivers like for instance those of ICOM (R71 etc.) and Kenwood (R1000
etc.), the input and IF bandpass filters were switched with diodes (N4148 or
equivalent). With a current of ~10 mA through a diode this works rather well,
regarding both IM and loss. But because at that time circuits weren't
simulated, it was overlooked that due to the diode capacitance, filters supposed
to be switched out, still exerted (often a nasty) influence. To eliminate that,
you'll also have to use a diode switch as a short circuit for the filter not in
use.

Jan

[bilou]

"garyr" <ga...@fidalgo.net> wrote in message
news:kkf2bi$uf5$1...@speranza.aioe.org...

Hello
IMHO diodes switches are by far the way to go.
They have been used for years in TV tuners at UHF frequencies.
They are very small and allow to place the switch exactly where it should
be.
They can support power as in TX/RX or att�nuator swithes.
Cmos IC's are used too in some SDR receivers but having to bring RF tracks
to a small SMD implies stray capacitance and inductance problems.



--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---

[bilou]

"Arid ace" <nons...@freedom.smokes.oil> wrote in message
news:pnpmm8d30s1qge5ro...@4ax.com...

But because at that time circuits weren't
> simulated, it was overlooked that due to the diode capacitance, filters
> supposed
> to be switched out,

Hi
Such effects do exist with any switches.
Behaviour of filters with a very mismathed input and output impedances is
very
unpredictable.
For switching IF filters logic would be to cascade them and shunt the too
narrow ones.
I have never seen this implemented.

[Arid ace]

On Wed, 17 Apr 2013 02:06:04 +0200, "bilou" <pad...@free.fr> wrote:

>
>"Arid ace" <nons...@freedom.smokes.oil> wrote in message
>news:pnpmm8d30s1qge5ro...@4ax.com...
> But because at that time circuits weren't
>> simulated, it was overlooked that due to the diode capacitance, filters
>> supposed
>> to be switched out,
>Hi
>Such effects do exist with any switches.

14p DIP reed relays (2 relays / package) already suggested by their
construction, to short circuit filters not in use, in order to prevent spurious
responses. That was in the 80s. Too expensive for mass production though.

>Behaviour of filters with a very mismathed input and output impedances is
>very
>unpredictable.

What matters is the spurious response via parasitic and other undesirable
capacitances or common inductances.

>For switching IF filters logic would be to cascade them and shunt the too
>narrow ones.
>I have never seen this implemented.

With each IF filter adding ~6 dB loss, either something has to be done about
gain correction or S-meter display. I built the 455 KHz IF with 2 filters for
each bandwidth and a low gain (~12 dB max) dual gate mosfet amplifier between
the filters. Switching bandwidth included removing the supply of the mosfets not
in use. The result was excellent - I verified that with a spectrum analyzer.
>

[K7ITM]

There are RF solid state switches designed for 50 ohm systems. Several companies have them: Hittite and Peregrine are a couple. At VHF and above they have very good third order intercept performance, but at least if the signals are fairly large, their intermod performance falls apart at lower frequencies. If you're operating below 0dBm, they may work well for you. They are tiny and easy to use...SPDT in a tiny 6-pin SMT pkg.

In an HF receiver I designed a few years ago, I tested many different RF switches. We needed something that would switch quickly (probably not an issue for you) and have very low distortion. The solid state switches I tried all had too much distortion, and I ended up selecting tiny reed relays (Meder SPST and Sanyu SPDT). They also contribute some distortion, but it was tolerable. They switch remarkably fast! The next best that I tried was Panasonic PhotoMOS "relays". They worked OK at 40M and lower frequencies, but had too much (non-linear) capacitance at higher frequencies. I considered diode switching using PIN diodes, but the net solution that way would have taken too much power (diode DC current), and PIN diodes with long enough storage times were pretty expensive.

You posted that you want to avoid using relays...I can appreciate that, but will still suggest that relays are a good way to switch RF with low distortion. Omron makes a nice DPDT RF relay that's quite small...the G6K-RF parts. Problem is that they are rather expensive. However, the same relay is available in a "non-RF" version, and its data sheet lists very good performance at lower RF frequencies...through 30MHz at least. They are relatively cheap, I think under $5 each in small quantities, and switch cleanly with practically no distortion. (I've tested for distortion with +20dBm input, and see distortion products around -140dBc...pretty much at my limit of ability to easily test. At that level, I'm not sure if the distortion is coming from the DUT or from the test setup.)

Cheers,
Tom

[garyr]

"K7ITM" <k7...@msn.com> wrote in message
news:bed39f7b-0d80-4247...@googlegroups.com...

I was trying to build an six-stage attenuator to place ahead of a bandpass
filter for a receiver. I tried using TS5A23159s, dual SPDT switches with
on-state resistance of 1 ohm. A bad idea and I lost my desire for an
attenuator for the present time. The cumulative effect of the switch
resistance and capacitance distorted the filter response; I have no
idea what the introduced distortion might have been. I had considered the
Omiron G6EK but opted to go with the low power solution. I was not aware of
the Hittite or Peregrine switches.

Thanks very much for your reply.
Gary