Nothern Utah WebSDR Logo - A skep with a Yagi Northern Utah WebSDR
A PTT mute for WebSDR use


What's this for?

It is common to use a WebSDR as an "auxiliary" receiver (or main receiver if your local noise is really bad!) and this means that when you transmit, your signal may be received by the WebSDR and get back into your transmit audio - usually via acoustical coupling - and cause an "echo" that both you and others listening may hear via your microphone.

Other than wearing headphones - which doesn't spare you from your own, delayed audio in your ears that can be very distracting - there are several ways to avoid this:
Figure 1:
A diagram of two circuits that can mute your computer's when you transmit to avoid the "echo".
Both versions include an "extra" contact that may be used to key an external amplifier, but this portion may
be omitted if you don't plan to use one.
In the vast majority of cases, the left-hand circuit will suffice:  The only difference between it and the right-
hand circuit is that the latter opens the ground path between the computer and the speaker which may
be useful of you hear some of your transmit audio getting into the speaker via RF pick-up.
Click on the image for a larger version.
PTT-triggered audio muting circuit
All of these require manual intervention and become cumbersome if you are engaged in quick and/or frequent band-and-forth, but there's another way:  Audio muting keyed by your transceiver.

The (simple!) circuit(s):

This circuits depicted in Figure 1 are very simple:  The audio from your computer to your speakers is routed through a set of normally closed relay contacts.  When the relay is energized by the radio doing into transmit, the contacts are opened, muting the audio from the computer.

The circuits depicted in Figure 1are "old school" using relays:  Relays were chosen because they are cheap, simple to use and offer good isolation at DC, audio and even RF.  Both versions show the keying of an external amplifier, but if you don't need to do that you can just leave that part out, needing just one DPDT (double-pole, double-throw) relay.

The components involved aren't at all critical:  It's recommended the small-ish relays with low-voltage coils (24 volts or DC or less) be used.  D1 is important as it suppresses the (high!) voltage from the relay coil(s) when they are de-energized and the magnetic field collapses:  Do not omit it!

Using with an amplifier:

The advantage of using relays is that it will worth with practically any amplifier, regardless of its keying voltage.  The only point of concern might be if you get a slight "pop" in the audio when you un-key the amplifier:  If your amplifier has a relay, it may be that its coil's electromagnetic field is collapsing and generating a high voltage/arcing across RY2's contacts and the a bit of this can couple into the computer audio:  The optional "snubber" network mentioned on the diagram should minimize this. 

Comments on radios' keying circuits:


While most "base station" radios have relays in them to key an external amplifier, some "portable" (usually QRP) radios use transistors instead, so it would be a good idea to check the voltage and current ratings on that output.  If its keying circuit can handle 15 volts or more and more than the current drawn by the relay(s) in the circuit, you are all set.

Note that transistor-based keying circuits are usually polarity-sensitive and ground-referenced (e.g. a positive signal on the external keying output gets "grounded" when you key the radio) so you would want to make sure that the keying line is connected appropriately (e.g. one side of the relays is tied to V+ and they are activated by "grounding" the other side of the coil.)


Another method of muting:  Cutting power to the speaker.

While it is preferred that one simply mutes the audio to the speakers, there are some powered computer speakers that can be susceptible to RF - that is, you hear distorted voice through the speakers when you transmit.  In many cases building the circuit on the right hand side of Figure 1 will take care of this as it completely disconnects the ground lead from the powered speakers, eliminating a possible RF current loop.  To find if this will work for you, try transmitting with the audio lead from the computer disconnected:  If it's quiet, when you transmit, this will probably work.

Some computer speakers are so sensitive to RF that this may not be enough and another method may be required:  Turning off the speakers' power when you transmit.

This would be done the same way as one interrupts the audio:  Wire a normally-closed contact from the relay in series with the DC power to your speakers.  This is easily done if your powered speakers use an external DC "power cube" - but if your speakers are high-powered or have a built-in mains (AC) operated power supply, this may not be either practical or safe.  In other words, this may not be applicable to all external "powered" speakers.

The down-side of cutting power to the speaker is that the audio may persist for a short time after the radio is keyed as its power supply capacitors discharge.  In addition to this, there may be a "click" or "thump" in the speaker as its power is removed and re-applied:  Whether this is objectionable depends on how badly your speakers might do this and whether or not it bothers you.

Powering the relay circuit:

The relay requires an external source of power - 11-15 volts being fine for a typical "12 volt" relay.  Out of convenience, the prototype was powered from a "wall wart" (power cube) that I had kicking around:  It's official rating is "9 volts DC" (unregulated) but because it consists only of a transformer, full-wave diode bridge and a capacitor, its voltage is around 12-15 volts under no/light load - perfect for this application.  Because the relay only consumes a few 10s of milliamps when keyed, the current rating of this wall transformer could be as low as 100 milliamps:  It does to have a capacitor-filtered output (but no need for regulation) or else the relay will chatter.  It is recommended that you NOT use a switching-type power supply for this as that will likely produce a bit of RF noise:  If you have a candidate power cube and it feels as though its box is empty (it has no "heft" to it) then it's likely a switching type.

If one side of the radio's PTT relay is (or can be) grounded (which is usually the case) it is possible to power this circuit from the same "12 volt" power supply as that which is running the radio.  If you do this, be sure to include a fuse (no more than 1 amp) in the power lead, near its connection to the rig's power supply to protect against accidental short-circuits.

Batteries may be used, too!

If you use 5 volt relays, it is also practical to power this circuit using four AA or AAA cells in series.  This has the advantage of providing a completely-isolated power supply and being able to avoid the inconvenience of yet another cable connecting to the circuit.  5 volt relays should easily be able to tolerate an extra volt or so from the "6 volts" of four alkaline cells in series (which happens only when they are fairly new) so there is no need to include a resistor.

A set of four AA alkaline cells should be capable of providing at least 50 hours of continuous key-down which means that if you are using these for SSB and you are often on the air, a set of batteries will last a year or so:  If "C" or "D" cells were used, the longevity would be about that of their shelf life!  If you run modes that require the transmitter to be keyed for long periods of time - like CW or some digital modes - you may not want to use battery power.

 
Figure 2:
Inside the audio muting box.  In the bottom-right corner is a "stereo" 3.5mm audio jack into which the external (powered) speaker plugs.  For convenience, all components were mounted and connected to a small piece of phenolic "prototype" board.
Click on the image for a larger version.
PTT-triggered audio muting circuit
"But I don't have any 12 volt relays handy!"

If you have different-voltage relays on hand, you can use those - just make sure that your power supply voltage is appropriate for them.   If all you have is 5 volt relays and your only convenient voltage source is 13.5 volts from your radio, you can use series dropping resistors, calculated as follows.
  1. Measure the resistance of your 5 volt relay's coil.  Let's assume for this example that your relay's coil measures 185 ohms.
  2. Calculate the current that the relay would draw at 5 volts.  Using Ohm's law, we divide the voltage by the resistance - so 5/185 = 0.027 amps (27 milliamps).
  3. Calculate the amount of voltage drop that you need by subtracting the rated relay coil voltage from our power supply voltage .  If we have 13.5 volts available, we need to drop (13.5 - 5 = )  8.5 volts.
  4. Calculate how many ohms it would take to pass the current calculated in step #2 at the voltage calculated in step #3.  In this case, we divide the voltage by the resistance, so ( 8.5 / 0.027 = ) 314 ohms.
  5. The two most common resistor values close to our example 314 ohms are 330 ohms and 270 ohms - but because we can get away with +/- 20%, either value will be fine.
  6. Now calculated the power dissipation in that resistor.  For this we multiply the current by the voltage drop, so we do (8.5 * 0.027 = ) 0.229 watts.  This is pretty near the maximum rating of a 1/4 watt resistor, so we would use a 1/2 watt resistor if we have one on-hand, but because it's being used only intermittently (e.g. while you are talking) a 1/4 watt resistor will probably be fine.
  7. Some relays (reed types, those with built-in diodes) can be polarity sensitive.  If the relay doesn't close when you apply voltage, try reversing the applied voltage's polarity.
To minimize the possibility of the relay's coil causing a "click" in the speaker audio when you unkey, it would be a good idea to put the reverse diode (D1 in the drawings) across the connection to the relay coil to suppress the transient voltage that would otherwise result.

If you are building the 2 relay version, unless both of your relays are identical, you'll want to do the calculation for each relay.  If both of your relays are identical, and they are both 5 volts you can wire them in series for 10 volts total, and then do the calculations for that, taking into account the fact that the coil resistance will also be doubled.  

An as-built example:

Here's an example of the "bare minimum" audio muting box depicted on the left side of Figure 1 (above).  Not only does it disconnect the audio going to the computer speaker, a third contact is used to key an amplifier.

The constructed unit can be seen in Figures 2 and 3.  A few notes on the construction:
Figure 3:
The outside of the muting box.  In the upper-left are the two cables with phono plugs:  One goes to the transceiver's keying line and the other connects to the amplifier.  On the right (the gray cable) is the audio from the computer and on this cable can be seen a ferrite tube that is used to minimize the possibility of RF being conducted on this line.  In the upper-right corner of the box is a black and white cable that goes to the filtered/unregulated transformer-type "wall wart" used to power the relay when keyed.
Click on the image for a larger version


PTT-triggered audio muting circuit
Conclusion:

When tested, the circuit described works perfectly.

One thing to remember is that any web-based receiver will have a slight delay as compared to a local, "real time" receiver meaning that when you un-key, you'll probably hear the last syllable or word that you transmitted coming back to you.  While this may be slightly distracting, echoes like this can never be heard over the air because this circuit guarantees that your speaker(s) will be muted any time that you key your radio.






Additional information:
 Back to the Northern Utah WebSDR landing page



This page last update 20181126.