Ok I'm not going to be a hypocrite. For years I have said that there is no reason to *buy* a soundcard interface. That statement remains true for the simple af and ptt switching designs.

The same is NOT true of the later designs like the microHAM microkeyer. These units are more flexible and offer complex usb to rs232 routing - they allow software control of audio routing and ptt timing. Yes you could build one yourself but these "ready made" designs are supported by the major logging programmes and integrate nicely. Yes, I now have two in use here.

That said, the following notes still apply if you just want a simple sound card interface

I wanted a simple interface to satisfy the following requirements:

Totally isolated AF interface
Totally isolated PTT and CW Key interface
Extend Accessory connector to interface to allow easy connection of TNC's etc
Provide a solid PTT or SEND connection that wouldn't be damaged by short cct's

All of my simple rig interfaces are homebrew , I can't see why anyone would want to buy a simple interface such as a rigblaster etc for a few hundred euros when one can be made from the typical junk box. Even if you have no junk box, the parts new couldn't cost more than 50 euro. You will see designs on the web all published by well meaning authors who claim to have designed their interface. Lets face it folks! a Darlington pair and a couple of transformers does not constitute a design. It's simply the use of well known building blocks to achieve an end. Dig out the veroboard and claim you designed it yourself if you like :-)

So point by point:

Isolated audio interface,

both TX and RX audio should be connected by audio transformers. Nothing complex about that but you need to pay attention to one detail. To avoid RFI and noise the interface should be built into an RF tight metal enclosure. Whatever connector you use to bring AF into or out of the interface enclosure must then be earth free. It's easy enough to mount the connectors on a small plastic panel but there is another issue to address. The outer screen of the cable from the PC soundcard would ideally have a fairly low impedance path to ground at RF ( but not 50hz ) so a fairly good low value capacitor would be connected from screen to the enclosure. Let's be a little ingenious here.. if we fix a piece of double sided copper clad board to the inside of the enclosure, one side is grounded and the fibreglass makes a fairly good dielectric. Thereffore the other side of the copper clad board is capacitively coupled to ground. There is of course no DC connection. Drilling suitable holes in the copper clad to mount the sockets and clearance holes in the box is a trivial matter. By etching away the copper as shown the screen connection of the sockets are only connected to our isolated and capacitively coupled ground. The connection to the soundcard side of the transformer is then completely isolated from the enclosure ground, but a low value capacitor couples the screen to the enclosure.

The radio side of the transformer is now connected to the line in and line out connections on the accessory connector. Checking the equipment manual for the correct pins and levels. Some radios will need an attenuator to drop the sound card level to the AF in level on the radio, Three resistors from the junk box. If you are fortunate enough to be connecting to an ICOM, and your PC soundcard has line in and line out connectors as well as the usual loudspeaker and microphone connectors, you will not need an attenuator. The radio fixed level rx AF is about -17dbm and you can set the line out level using the soundcard mixer to about -17dbm. OK one last point about the interface, the soundcard output is generally stereo, a left and right signal. You will either need to terminate the unused line into a resistor ( see diagram ) or make it available for your headphones. Some DSP software lets you listen to the processed audio on one channel while generating, for example HSCW, on the other. If you make use of this audio you MUST maintain the screen isolation. Don't make the mistake of connecting it to your SO2R box with a patch lead then wonder where the hum is from.

Isolated PTT and CW Key interface,

Most software uses the RS232 signal lines RTS for PTT and DTR for CW Key. The levels offered by most modern RS232 ports switch between a few volts negative to a few volts positive when operating. Its a simple matter of a diode or two for transient protection and a small resistor to limit the current and hey presto we have enough current to safely light an LED. Fantastic! :-) if that LED is within an opto isolator such as a 4n23 or whatever else you have in your junk box we now have a transistor that will sink about 50mA under software control. This transistor isn't connected to the rs232 ground pin.. so it's isolated, about 7500V worth of isolation if the right opto isolater is used. Now that means that you need to make sure the board layout is sensible, keep the rs232 connections on one side, the ptt and cw key connections a the other, use insulated hookup wire etc. All simple straight forward common sense.

So what to do with the output from the opto isolator?

The connection to the CW key line on an IC746 sinks about 4mA on key down with a key up voltage of about 8v. So with the inclusion of a transient protection diode we are fairly safe to connect that straight to the opto isolator as shown in the diagram. If you are keying something like an FT225, you need to sink a bit more current, the use of a darlington pair as I am about to describe for the PTT interface will be required

The PTT line is a little more involved. First of all, the PTT line on the IC746 needs to sink a bit more current. There are also two of them! on the first accessory connector there is the normal HSEND line. The VSEND line is on the second accessory connector. Why Icom didn't make these two lines software configurable is a mystery to me... well not really, none of the manufacturers (KenComAesu) expect radio amateurs to connect anything other than a "genuine" accessory to their interface ports.

So first of all you will need to provide a bigger transistor to do the switch, simple enough just use a Darlington pair. Connection is simple as per the diagram. This will pull the HSEND or VSEND line toward ground when the opto is "lit up".
When I say "toward" ground this raises an important point. The Collector will sit at around 600mV and this is enough to reliably make the radio got to TX. However if you recall we have two PTT lines, one for HF and one for VHF. Now you can simply connect both lines together as a crude "get around" or you could provide a switch to direct the PTT signal to the correct line. But if you do this you lose some flexibility. It's quite possible that you have a big VHF amp and a separate HF amp. It's also probable that they are normally "on" providing a source of hot air to keep the shack and the cat warm, you have to make sure you are ready at a moments notice to work the DX. You don't want the HF amp biasing up every time you tx 2m meteor scatter so you need to isolate the two lines. If you try to switch the lines by using a pair of diodes you will not be able to get either line close enough to 0V to bring the radio to TX, the collector of the key transistor is at 600mV and you will lose another 600mV ( ish ) in the diode. The line will sit at about 1.2V and the radio will ignore it. I leave it to your imagination how to get round this, but you can either do the isolation with a couple of diodes after the opto isolator and provide two keying transistors, you can use two opto isolators and place the LED's in series, You can use a VMOS transistor which has a very low on resistance, then the 2 diodes will do the trick. All of the above will work, I used VMOS 'cos I have them in the junk box.

Incidentally, my IC756pro2 will key straight from a 4n23.. not needing anywhere near the HSEND current of the IC746.

Extend accessory connector

If you connect multimode TNC's, test equipment, SO2R boxes etc to your radios you will probably spend a little time each week grumbling about having to find the connectors at the back of the radio. Put a connector on the front of your interface, direct the various interface lines such as PTT and TX AF through a front panel switch and you can save all of that effort. It will take you an extra hour to build and save many hours of struggling in the future.

Solid PTT output

If you found your way to this page from my IC746 notes, you will already know that the SEND relay in the IC746 is fairly fragile. It's also a relay and hence quite slow. When you hit the PTT or press the paddle the SEND relay operates some 10mS later, the RF is arriving at the ANT connector only a few mS after that. If you use the SEND relay to control a big amp, a few relays and an RX pre amp this should worry you! If you sequence your station ... you do don't you? you need a fast control line from the radio to the sequencer. The diagram shows a way of using a Hexfet or a Vmos transistor to provide a fast switch to 0V on PTT. At worst you can use this to key your HF amp, at best it will provide the necessary interface to your station controller. If you recall from the IC746 notes on transverter interfacing, the relay used to direct RF to the transverter port can be controlled by the station controller to delay RF from the radio until the sequencer tells it that it's safe to TX. No more worries about accidentally using break in, MOX or VOX and inadvertently bypassing your sequencer. The interface can also be used to switch the transverter to TX... adding the circuit will add about an hour to the construction and adds flexibility.

Schematics

The following diagrams and photos should make all of the above clear, It's not a construction article.. just a collection of ideas. Use them as building blocks and make the interface section by section. Go on get the iron out.. save yourself 200 euros. interface part 1 shows the basic isolated af ptt and cw key schematic