I'm looking for a passive T/R switch for a noise canceller project.
The idea is to BYPASS the noise canceller circuit AUTOMATICALLY and
IMMEDIATELY upon transmitting, even when the noise canceller is turned
off. This is VERY important, because transmitting power into the
noise canceller could fry some of its components.

Are there places to buy T/R switches? Or must I build one myself?

Jason Hsu, AG4DG

On 20 Sep 2003 08:25:56 -0700, (Jason Hsu)
wrote:

I'm looking for a passive T/R switch for a noise canceller project.
The idea is to BYPASS the noise canceller circuit AUTOMATICALLY and
IMMEDIATELY upon transmitting, even when the noise canceller is turned
off. This is VERY important, because transmitting power into the
noise canceller could fry some of its components.

Are there places to buy T/R switches? Or must I build one myself?

Jason Hsu, AG4DG

Jason:

How fast to you want it to operate? If you can tolerate a few mSec
(say 7 or 8 mSec after bounces cease) operate/release time, an
inexpensive Omron G2RL-24 relay will do quite nicely at the 100 watt
level.

I'm working with N8LP on an article characterizing those relays as RF
switching devices, and so far we are quite pleased with the measured
results.

And, of course, an advantage of a relay is that you can wire it so
that when the canceller is unpowered, the contacts switch the device
out of the circuit, so it meets your requirement.

If you can't tolerate the time to operate a relay, then a PIN diode
switch may be your answer. Not cheap and not necessarily easy, but
FAST. If you can accept a few mSec delay, but not 7 or 8 mSec, then
you can speed up a relay operate/release times, or try a faster relay,
such as some of the reed devices that operate in 1 mSec or
thereabouts.

If you are at QRP power levels, then ordinary signal diodes may work
in a diode switch. It may not meet your power-off requirement, though.
Of course, a combination relay for power-off protection and a diode
switch for normal activity could do it.

As far as buying a T/R switch, relays are certainly available
(Dow-Key, for example). Since almost everyone these days uses a
transceiver, the market for stand-alone T/R switches has greatly
diminished since the days of separate TX and RX.


Why don't you define your requirements a bit for us, such as:
(1) power to be switched
(2) speed for switch to operate/release
(3) power available to operate the switch
(4) cost target
(5) frequency range (assumed to be 3-30 MHz since you are working on a
noise canceller)
(6) Isolation between TX and RX port
(7) Maximum acceptable loss


Jack K8ZOA

On 20 Sep 2003 08:25:56 -0700, (Jason Hsu)
wrote:

I'm looking for a passive T/R switch for a noise canceller project.
The idea is to BYPASS the noise canceller circuit AUTOMATICALLY and
IMMEDIATELY upon transmitting, even when the noise canceller is turned
off. This is VERY important, because transmitting power into the
noise canceller could fry some of its components.

Are there places to buy T/R switches? Or must I build one myself?

Jason Hsu, AG4DG

Jason:

How fast to you want it to operate? If you can tolerate a few mSec
(say 7 or 8 mSec after bounces cease) operate/release time, an
inexpensive Omron G2RL-24 relay will do quite nicely at the 100 watt
level.

I'm working with N8LP on an article characterizing those relays as RF
switching devices, and so far we are quite pleased with the measured
results.

And, of course, an advantage of a relay is that you can wire it so
that when the canceller is unpowered, the contacts switch the device
out of the circuit, so it meets your requirement.

If you can't tolerate the time to operate a relay, then a PIN diode
switch may be your answer. Not cheap and not necessarily easy, but
FAST. If you can accept a few mSec delay, but not 7 or 8 mSec, then
you can speed up a relay operate/release times, or try a faster relay,
such as some of the reed devices that operate in 1 mSec or
thereabouts.

If you are at QRP power levels, then ordinary signal diodes may work
in a diode switch. It may not meet your power-off requirement, though.
Of course, a combination relay for power-off protection and a diode
switch for normal activity could do it.

As far as buying a T/R switch, relays are certainly available
(Dow-Key, for example). Since almost everyone these days uses a
transceiver, the market for stand-alone T/R switches has greatly
diminished since the days of separate TX and RX.


Why don't you define your requirements a bit for us, such as:
(1) power to be switched
(2) speed for switch to operate/release
(3) power available to operate the switch
(4) cost target
(5) frequency range (assumed to be 3-30 MHz since you are working on a
noise canceller)
(6) Isolation between TX and RX port
(7) Maximum acceptable loss


Jack K8ZOA

Jack Smith wrote in message . ..

Why don't you define your requirements a bit for us, such as:
(1) power to be switched
(2) speed for switch to operate/release
(3) power available to operate the switch
(4) cost target
(5) frequency range (assumed to be 3-30 MHz since you are working on a
noise canceller)
(6) Isolation between TX and RX port
(7) Maximum acceptable loss

What are these parameters for the T/R switches in the MFJ-1025/1026
noise canceller and the T/R switches in the receiver circuits of
transceivers? The advertising for the MFJ noise canceller explicitly
says that it uses an RF sense T/R switch. I'm sure that transceivers
need T/R switches to prevent the RF from entering the receiver. That
can't possibly be good, because receivers are designed to be very
sensitive, NOT handle large amounts of power.

Jason Hsu, AG4DG

Jack Smith wrote in message . ..

Why don't you define your requirements a bit for us, such as:
(1) power to be switched
(2) speed for switch to operate/release
(3) power available to operate the switch
(4) cost target
(5) frequency range (assumed to be 3-30 MHz since you are working on a
noise canceller)
(6) Isolation between TX and RX port
(7) Maximum acceptable loss

What are these parameters for the T/R switches in the MFJ-1025/1026
noise canceller and the T/R switches in the receiver circuits of
transceivers? The advertising for the MFJ noise canceller explicitly
says that it uses an RF sense T/R switch. I'm sure that transceivers
need T/R switches to prevent the RF from entering the receiver. That
can't possibly be good, because receivers are designed to be very
sensitive, NOT handle large amounts of power.

Jason Hsu, AG4DG

On 20 Sep 2003 19:15:23 -0700, (Jason Hsu)
wrote:

Jack Smith wrote in message . ..

Why don't you define your requirements a bit for us, such as:
(1) power to be switched
(2) speed for switch to operate/release
(3) power available to operate the switch
(4) cost target
(5) frequency range (assumed to be 3-30 MHz since you are working on a
noise canceller)
(6) Isolation between TX and RX port
(7) Maximum acceptable loss

What are these parameters for the T/R switches in the MFJ-1025/1026
noise canceller and the T/R switches in the receiver circuits of
transceivers? The advertising for the MFJ noise canceller explicitly
says that it uses an RF sense T/R switch. I'm sure that transceivers
need T/R switches to prevent the RF from entering the receiver. That
can't possibly be good, because receivers are designed to be very
sensitive, NOT handle large amounts of power.

Jason Hsu, AG4DG


Jason:

A T/R relay in a transceiver has an easier job of it, because the
designer can sequence the operation, along the following lines:

Key Down -- Start TR Relay in motion -- wait a few mSec -- turn TX on
and start putting out power.

By sequencing the operation, you have two advantages (a) the relay
doesn't have to hot switch RF and (b) the relay doesn't have to be
blindingly fast, a few mSec is fine.

Your case is different -- you will have to sense the RF and hot switch
it. Still, a relay should do what you want.

I have a JPS noise canceller and it uses a simple relay. I suspect
something similar would work for you. I can recommend an Omron
G2RL-24 relay. You will be able to obtain at least 40 dB isolation up
to 50 MHz, and closer to 75 dB below 10 MHz. That's more than enough
to protect the front end. And, we have put well over 1x10^6 operations
on one hot switching switching 100 watts RF without failure. In fact,
I've got close to 100,000 operations switching 400 watts without a
failure so far.


Jack K8ZOA

On 20 Sep 2003 19:15:23 -0700, (Jason Hsu)
wrote:

Jack Smith wrote in message . ..

Why don't you define your requirements a bit for us, such as:
(1) power to be switched
(2) speed for switch to operate/release
(3) power available to operate the switch
(4) cost target
(5) frequency range (assumed to be 3-30 MHz since you are working on a
noise canceller)
(6) Isolation between TX and RX port
(7) Maximum acceptable loss

What are these parameters for the T/R switches in the MFJ-1025/1026
noise canceller and the T/R switches in the receiver circuits of
transceivers? The advertising for the MFJ noise canceller explicitly
says that it uses an RF sense T/R switch. I'm sure that transceivers
need T/R switches to prevent the RF from entering the receiver. That
can't possibly be good, because receivers are designed to be very
sensitive, NOT handle large amounts of power.

Jason Hsu, AG4DG


Jason:

A T/R relay in a transceiver has an easier job of it, because the
designer can sequence the operation, along the following lines:

Key Down -- Start TR Relay in motion -- wait a few mSec -- turn TX on
and start putting out power.

By sequencing the operation, you have two advantages (a) the relay
doesn't have to hot switch RF and (b) the relay doesn't have to be
blindingly fast, a few mSec is fine.

Your case is different -- you will have to sense the RF and hot switch
it. Still, a relay should do what you want.

I have a JPS noise canceller and it uses a simple relay. I suspect
something similar would work for you. I can recommend an Omron
G2RL-24 relay. You will be able to obtain at least 40 dB isolation up
to 50 MHz, and closer to 75 dB below 10 MHz. That's more than enough
to protect the front end. And, we have put well over 1x10^6 operations
on one hot switching switching 100 watts RF without failure. In fact,
I've got close to 100,000 operations switching 400 watts without a
failure so far.


Jack K8ZOA