Print

[N3KXZ]

I got this cheap coaxial antenna switch at a hamfest, thinking to use it inside the shack to compare different antennas and gear. My intention was to turn it around and use different radios for the same antenna- for instance, I have been comparing the reception (and features) between different rigs using the same antenna. I figured these things are just switches and they should be bidirectional, right?

Well, last night I had two rigs hooked into it on two of the "outputs" (only for receive)  in order to switch between them using my one antenna, and found that I got bleed through! In other words, I would select one rig but still faintly hear the other rig's reception coming through- the unselected rig was still receiving through the antenna (which I could hear when turning the volume on the selected rig all the way down). Uh, no. So now I don't trust this cheap switcher at all for anything. Glad I wasn't attempting to transmit! I was using this at HF reception frequencies, and of course the outside coax conductor is grounded.

So, how could it be that this thing is not bidirectional? Or is the issue simply that this particular switch is so cheaply made. Isn't is common for these types of coax switchers to be bidirectional? Do you have suggestions on how to evaluate coax switches for future purchase?

Thanks,
Keith
N3KXZ

[W1VT]

Someone neglected to shield the two RF paths from each other.

[GEORGEMINK]

If you want to take the time and rewire it using shielded cable that would help, grounding the shield on all of the jumpers.
I have MJF antenna switches and use them just as you intended and have never had any issues.

[N3KXZ]

Thanks for that idea George, I might do that.

[WD8OTT]

What you are experiencing is poor port to port isolation. It is usually measured in decibels or db's for short. The more db's of isolation, the better. This switch doesn't have very good port to port isolation. That being said, if you try to use it as you did, when you transmit, the transmitted signal will go out to the antenna but also back feed into the other transceiver. Definitely a BAD THING. You will also notice two wires on each center pin of the connectors. That is because the switch contacts cannot handle much power. That actual switch body is two switches in parallel. Each handling part of the current. With the wires going all over the place willy nilly the switch probably has an impedance bump and isn't a constant 50 ohms. To some it up, that switch could be used to switch multiple antennas to one RECEIVER. Possibly switch multiple antennas to a transceiver of LOW POWER. Personally, I'd get rid of it.

[K8AC]

Keith - the switch may be cheap, but I assure you that it IS bi-directional.  If you used the switch to connect multiple antennas to a single transceiver, the transmit signal would be coupled to the non-selected output ports, but at a greatly reduced level.  The better switches, such as an Alpha-Delta, have internal construction that minimizes the capacitive and inductive coupling between ports, but do not eliminate it.  The signal coupled to one of the non-selected ports may be 60-80 dB down, but if you're running a KW through the switch you'll still be coupling a detectable signal to those ports.  With the construction of this switch, you can be sure the signal coupled to the non-selected ports will be substantial and possibly enough to do damage to the front ends of other transceivers connected to the switch.  Even the MFJ switches are built to try and minimize the unwanted coupling. 

[N2SR]

What you are experiencing is poor port to port isolation. It is usually measured in decibels or db's for short. The more db's of isolation, the better. This switch doesn't have very good port to port isolation. That being said, if you try to use it as you did, when you transmit, the transmitted signal will go out to the antenna but also back feed into the other transceiver. Definitely a BAD THING. You will also notice two wires on each center pin of the connectors. That is because the switch contacts cannot handle much power. That actual switch body is two switches in parallel. Each handling part of the current. With the wires going all over the place willy nilly the switch probably has an impedance bump and isn't a constant 50 ohms. To some it up, that switch could be used to switch multiple antennas to one RECEIVER. Possibly switch multiple antennas to a transceiver of LOW POWER. Personally, I'd get rid of it.

Decibel is abbreviated dB, not db. 

[W9IQ]

Someone neglected to shield the two RF paths from each other.

I am presuming the housing is metal so that would complete the ground path on the connectors.

But the general description for that design is "cheap". The isolation will be poor as a result. If you use it to switch two transceivers, it may cause damage.

- Glenn W9IQ

[K0UA]

I have that exact same switch in one of my junkboxes around here.  Would you like another?

It is just garbage, and you cannot expect much performance from it.

[N3KXZ]

I wonder... what would be a simple and safe way to test/quantify the port-to-port isolation? One way that comes to my mind is to connect an inline wattmeter and dummy load to each of two of the ports and watch the power going out of the unselected port as I transmit. Is there an easier method? Maybe a way to use the nanoVNA? Or?

I ask because I might want to play around with improving the switch by shielding and rerouting the cables, and maybe eliminating one of the parallel switches (I only need 100 watts maximum). Might be fun and educational if I had an easy way to test the port isolation.

Of course the best thing to do is just throw this away and buy an Alpha Delta switch, but I have more time than money...

[WB6BYU]

You can measure it with a 2-port VNA.  Put a dummy load on
the common port of the switch, and connect the VNA to a pair
of adjacent ports.  The measured path loss should give you
the isolation between the two ports.

(You probably want to set the switch to VNA PORT 1, so it
has the termination on it.)

Otherwise, make a simple RF voltmeter probe using a diode
and connect it to a digital voltmeter.  Connect that to the
unselected port while transmitting at a fixed power level
through the switch into an antenna or dummy load.  That
will measure the induced voltage on the unsellected port.
Might be more accurate to include some sort of load resistor
as part of the RF voltmeter.

[AI5BC]

You can measure it with a 2-port VNA.  Put a dummy load on
the common port of the switch, and connect the VNA to a pair
of adjacent ports.  The measured path loss should give you
the isolation between the two ports.

But why waste the time? You can tell immediately it is designed and built by an amateur who has no clue how to design a coax switch. There is no way a mess of wires connected to a rotary wafer switch is going to pass anything more than DC and AC power. All it will tell you is the box it came in is worth more than the switch.

OP go search for Dowkey 78 series switches. They are military/laboratory grade coax switches. 1 Kw of power up to 450 MHz with 60 dB isolation Better than anything you can buy in a ham schack They ar eno longer made but a lot of them on the after-sale marker like RF Parts has several. Here is a link to the product sheet so you can search for your specific application. a good 3P3T for around $100

http://socal.mavin.com/pictures/dow-key.pdf

[K6AER]

Actually the switch is fine if all you are doing is selection antennas. 30 dB of isolation is fine.

Better antenna switches ground the unused ports and provide up to 50 dB of isolation such as the Alpha Delta series. When in the Com. off position all antenna ports are grounded. The problem when using your  antenna switch to select transceivers you need much better port to port isolation to protect receiver front ends.

[ON6AB]

If you're talking about a cheap switch, the Daiwa CS201 immediately comes to mind.
By coincidence, I justed measured the losses and isolation of this switch with a NanoVNA.
results are:
1- LOSSES:
- 50Mhz: 0,01dB
- 144Mhz: 0,01dB
- 430Mhz: 0,15dB
These losses are so low that you can argue if they're caused by the switch or by inacurate measuring.
In any case, they are neglectable (up to 144Mhz).

2- ISOLATION:
- 50Mhz: 74dB
- 144Mhz: 60dB
- 430Mhz: 50dB

These measurements show that although the results are a bit worse than the DAIWA specifications, this switch is perfectly usable on HF and up to 144Mhz.
Bear in mind that  strong signals on HF can easily be around -30dBm and receivers are capable of hearing signals of -120dBm without any problem.
That's a 90dB difference! With a - perfectly adequate- isolation of 70dB, Of course you will hear them on the other port.

Can you damage your rig by sending 100W on one port with another rig on the other one?
No, 100W is 50dBm. With 70dB isolation, only -20dBm of this signal is presented on the other port which is strong, put perfectly acceptable for any receiver.

[VE7RF]

IF the Un-used portS are DC grounded to the chassis of the switchbox, I woulda thought the isolation would be through the roof. Like, just how are you going to hear a signal, any signal, when the center conductor of the UN-used port  is DC grounded ?

That's like putting a shorted PL-259 plug into the back of xcvr #2...with ant connected to the SO-239 of xcvr #1.  You won't hear anything on  xcvr #2.

Another possible option is to try using 2 x ports that are NOT adjacent to each other.... like say ports  #1 and  #3...... or ports  #1  and #4.