Sound Card<–>Radio Audio Isolation

[VK2CSW]

Hi All,

A noob question, if I may.

I am (finally) cobbling some gear together to have a play with this FT8 business.

As the PC and the radio are powered from different sources, one of which is a switch mode power supply, I gather that it would be prudent to have some form of isolation between the radio and the sound card.

Would a pair of 1:1 transformers be enough?  Or would optocouplers be a better solution?

[VK4ADC]

Would a pair of 1:1 transformers be enough?  Or would optocouplers be a better solution?

The 1:1 transformers like the 600:600 ohms versions available around the place  (Jaycar MA1512, though their MM5534 3K:3K is probably a better choice) are fine.   One for receive audio, one for send.
I have a homebrew interface that I am presently using for FT8 on HF and I have included potentiometers on the output sides (pot values not critical - 5K etc) so that I can set audio levels without resorting to changing any mixer settings in Windows.  They are connected through to the Accessories port on the radio and it gives constant level on receive and a level independent of the mic gain setting on transmit (Icom radio).

Optocouplers (ie LED diode + phototransistor device) are for DC-based signalling (on/off) and won't operate with direct audio.

[VK2CSW]

Hey Doug,

Much appreciated.

Time to rat through ye olde PC parts box to see if I've got any old modem cards left, from which to scavenge some transformers.

Thanks

[VK2CSW]

Well, I knocked up an interface this afternoon.

That was the easy bit.  Quite soon frustration set in.  First up WSJT-X won't talk to the radio via the CAT cable. (I suspect the cable - cheap Chinese fake prolific job.).  I have what I hope is an FTDI one on the way from the UK.

I am using an external USB sound card. At first it didn't want to play nice, then suddenly around sundown I started getting decodes on 20m.  Pretty exciting, even if I couldn't reply.

Then the PC decided to randomly switch back to using the internal microphone, despite the software wanting to use the USB card.

After mucking around for a couple of hours, the waterfall looks like it is receiving but no more decodes - of course it may be that there was nothing to decode after sunset.

[VK4ADC]

Colin

If you dont want to wait for the shipment from the UK, try this Arduino USB adapter from Jaycar : https://www.jaycar.com.au/arduino-compat...e/p/XC4464
"A mini-USB to 6-pin serial port module used to communicate with Arduino boards and modules. It uses the original FT232 chip with power/sending/receiving indicator and also a selectable 3.3V and 5V power option to suit different boards."  A link to the Windows driver is on the packaging (well it was on the one I bought a while back).
Win7 is ok with it here on several different computers.

For Icom CIV:  Fit a 100 ohm resistor between TXD and RXD pins, wire from RXD to the centre connector pin of a 3.5mm plug, outer of the plug back to the adapter's ground pin, preferably in a single core SHIELDED audio cable (not Fig8).    Use shielded wire to prevent CI-V comms issues with RF presence. I also run the cable a few turns through iron-powder toroids or use ferrite clip-ons to reduce RF noises back into the receiver with CI-V "chatter".

For Yaesu CAT, use twin shield audio cable back to the transceiver connector from the TXD and RXD pins... 

I am using a cheapex USB-sound adapter from eBay also and no sign of the issue you describe. It sounds like you have a bad USB lead if it drops back or bad port connectors if it drops back to the internal sound card intermittently.

eg..  "USB 2.0 to 3D AUDIO SOUND CARD EXTERNAL ADAPTER VIRTUAL 5.1 CH MIC Headphone"
eg https://www.ebay.com.au/itm/USB-2-0-to-3...0951090739

Have fun on FT8 

[VK2CSW]

Coincidentally, I ordered that very interface last night.

After posting my last post, I decided the try a different tack and (again) re-installed the old driver and then tried connecting to the radio via the CAT cable using CHIRP. After a couple of attempts, which is usual with CHIRP, it connected and downloaded the data.

I then ran up WSJT-X and it connected and took control of the radio. Happy Happy Joy Joy!

After some physical fiddling with the sound card and connecting it via a USB hub it started decoding again.

So close.

The radio (a Yaesu FT-857D by the by) was switching to transmit where appropriate but the on board vswr display stayed on zero, so I think it indicates that nothing much in the way of RF is going out of the radio.

By then it was a bit late to dig further - although it did make me wish I still had access to a CRO. I'd like to see if or how much audio is coming out of the sound card as the next diagnostic step. In lieu of a CRO I will clip some headphones across the output of the isolation transformer and see what can be heard.

Getting much closer. Unfortunately I am away for work for a week or so, meaning the chase will have to wait.

[VK2CSW]

Sometimes the solution is in a lateral direction.

Right near the radio in question is a bunch of music kit, including a small mixing desk that will talk to the PC via USB. Allowing me to eliminate the external sound card.

When I get back I'll route the audio I/O from the data port via the isolator board into the tape in/tape out channels. That way I can push the input and/or output to somewhere I can hear it. I can also control the gain of the signals. Then I 'll actually have some idea what is going on.

[VK5ZD]

Just a thought. I think the FT-857/897 ignores the audio from the data connector unless you select DIG mode.
There is also a digital VOX option that will put the radio into transmit when audio appears. Saves messing about with the CAT cable.

[VK4ADC]

Just a thought. I think the FT-857/897 ignores the audio from the data connector unless you select DIG mode.
There is also a digital VOX option that will put the radio into transmit when audio appears. Saves messing about with the CAT cable.

Iain

The CAT connection is still desirable as the WSJT-X software sends band/freq data back to PSKReporter if that tickbox is ticked (and it generally is if the PC has an active internet connection available). Without CAT, it can't be sure what band and frequency the radio is actually tuned to.

Band change is actually done from a drop-down box in WSJT-X and it sets the radio's frequency for the specific mode eg FT8 dial is different to that for JT9, WSPR, etc..


73 Doug VK4ADC

[VK2CSW]

Wow. What an afternoon.

Talk about frustrating. While I was away the bits I ordered turned up.

So full of anticipation I plugged in the all new $3.50 USB sound card and $40 CAT Cable.

First up, what was supposed to be an FTDI chip in the CAT cable, turned out to be a Prolific and probably a fake as I needed to load some old drivers and run it up using CHIRP first. It mostly works.

Which is more than I can say for the sound card. It is worse than the one it was to replace. Nada, Zilch, Nothing.

OK. I break out the USB mixing desk. I plug in the phones, run up some mixing software and all good it works, both in and out.

I run up the FT-857 with data I/O through the isolation board I built and via the desk and lo and behold, decodes all over the place. So far so good.

Still no output. The radio is triggering but nothing is going out. Checked a few settings on the radio with no joy.

Now the fun begins. For tracing purposes I decided to use the Phones Out of the desk to drive into my interface board. Using crocodile clips and an old speaker I started tracing the sound.

It doesn't get to the end of the Data I/O cable. Nor does it get to the output of the board. Nothing at the output of the transformer. I jumped to the input of the board and there is sound.

I then bypassed the small filter capacitor with a jumper and now I get the sound at the input of the transformer, but still nothing at the output.

On a whim, I reverse the transformer still nothing. Looks like we have the culprit. I then turned the Phones fader right up. I was getting the smallest signal on the output.

I fed sound into the other transformer and the signal on the output was somewhat attenuated but nothing like the other one.

The tally so far is two faulty USB Sounds Cards, two bodgey CAT cables, a possibly faulty capacitor, a faulty transformer and a noisy USB hub power supply.

Someone doesn't want me to get this working.

Next step will be to get some 1/4" plugs and make up a Data I/O lead to allow me to bypass the isolation board and test via the desk and maybe a DI box to break any ground loops. Actually, I am amazed I didn't have any 1/4" plugs kicking around.

Onwards and Upwards (and backwards and sideways).

[VK3ALB]

I have purchased a number of AXE-027 CAT cables from Altronics and although relatively expensive compared to ebay offerings, found them to be very good.

The drivers install and work without hassle and of all the different USB cables I have tried I found these radiate the least amount of noise. They're not perfect but even when used directly beneath high gain antennas the noise is within acceptable limits.

[VK2CSW]

Thanks Lou,

I'll have a look at them.

To round it off, the USB mixing desk took a little coaxing to get it talking to WSJT-X and then to work out fader, gain and EQ.  For the record EQ I am using is 12kHz +15 (as far up as it will go) and 80Hz -15.  The Rx VU level on the desk is around 0 to +6 dBu.

I fiddle with the mic gain, input fader to get the Rx working.  I use the Main Mix Out fader to set the Tx signal below clip - again around +6 dBu.  Of course it is in combination with the PWR control in the software.

I completed a couple of QSO's with some JA's but as my location and antennas leave a little to be desired, that is a good start.

Thanks for the help, it has been much appreciated.

   

[VK2CSW]

I spent a few frustrating hours **** around with the interface board I had knocked up. While the system works with the mixing board, sometimes RF gets into it and it goes crazy.

I appears that the headphones I was using to monitor things are themselves an antenna. A few turns through a toroid ferrite helped, until the cat decided to help and knocked the headphones off the bench, shattering the ferrite...

So still no joy with the DIY interface.

I have pared it back to a DIN connector, some cable, vero board with two 3k Ohm 1:1 transformers, some more cable, 3.5 mm connectors.

The Rx signal is quite low but gets from one end to the other. The Tx gets to the board, but not to the radio. I even soldered jumpers to bypass the transformer with no luck.

All that is left to swap out is the DIN connector and I don't have one to hand, a visit to Jaycar is in my future I guess.

Normally, at this juncture, I'd borrow a CRO and signal generator from work and start tracing. Unfortunately where I work these days doesn't have such things. (an unforeseen downside to moving to a less technical role I suppose).

Anyhoo, it has 'forced' me to order one of the little TFT LCD Oscilloscope kits from fleabay (https://www.ebay.com.au/itm/282873100082) for about $30. I have a couple of audio projects on the workbench, so it might be useful elsewhere. Other than replacing the DIN connector, I guess this task is on hold until I get and build the CRO.

[VK3ALB]

I have purchased a number of AXE-027 CAT cables from Altronics and although relatively expensive compared to ebay offerings, found them to be very good.

The drivers install and work without hassle and of all the different USB cables I have tried I found these radiate the least amount of noise. They're not perfect but even when used directly beneath high gain antennas the noise is within acceptable limits.

Yeah I better revise my suggestion for these cables. They are good BUT they use inverted logic. This means they won't work directly as you expect. They work for me because I use a picaxe where I invert the logic on the code. I was reminded of this today during a particularly frustrating troubleshooting episode.

[VK2CSW]

Further progress, then a further setback (why do I do this to myself?)

I decided to rebuild the interface from scratch.

It just the 3 connectors and 2 transformers on some veroboard.

I ordered a DSO138 2.4" TFT 200 kHz oscilloscope from ebay which has made things easier to trace (I mean how hard can it be?).

I had an intermittant Rx issue which I was able to track it down to the 3.5 mm TRS connector I was using on the audio in to the PC, the length of the sleeve and ring did not match the dodgy TRRS to 2x TRS adapter I was using.

I replaced the TRS plug with a less dodgy one and voila Rx was back.

Now for Tx. The 'scope allowed my to follow the audio path from the radio end back towards the PC. Two issues presented.
1. A cracked track on the verocard - bit of solder and it was fixed.
2. A massive loss of signal via the transformer. I had already replaced this transformer. A head scratch or two and I decided to refit the original transformer.

Now Tx works!

I then put some pins on the veroboard so I can use the 'scope to monitor the shape of my outgoing signal.

Time to chase some contacts...

But no. The Mrs comes in a says the soundbar on the TV is going nuts - volume going up and down, changing settings etc etc etc.

Checked the obvious - VSWR is 1.5:1, my transmissions were being received so I doubt it was being overdriven, station earth is intact and bonded to mains earth.

The soundbar is linked to the TV via an optical cable, so I doubt that is the point of ingress. I placed some clip on ferrites on the power lead with no change.

Time for more head scratching.

[VK4ADC]

Further progress, then a further setback (why do I do this to myself?)
....
But no.  The Mrs comes in a says the soundbar on the TV is going nuts - volume going up and down, changing settings etc etc etc.

Checked the obvious - VSWR is 1.5:1, my transmissions were being received so I doubt it was being overdriven, station earth is intact and bonded to mains earth.

The soundbar is linked to the TV via an optical cable, so I doubt that is the point of ingress.  I placed some clip on ferrites on the power lead with no change.

Time for more head scratching.

Colin

It is likely that the soundbar, despite being linked by an optical cable, is suffering what we used to call audio rectification issues - in this case rectifying the RF while on TX and causing DC level changes inside.  Does the effect happen regardless of the band you are transmitting on - or is the antenna in near proximity to the sound bar ? Is the power lead from a power adapter acting as an antenna (a few turns through a clip-on ferrite will prove this aspect) ??  This type of problem is not unusual but at least the problem is within your household so therefore resolve-able without bad neighbour issues.

I have an occasional issue on 20M with FT8 only when I have the triband yagi pointing over the top of the shack and 100W out on transmit.  Fortunately that direction is NNE and I seldom have to point it that way.

My issue is with the CAT control (CI-V actually since it is an Icom) of the radio. The frequency intermittently jumps because the CAT data is being corrupted despite using shielded CAT wiring.  Extra ferrites haven't made any real difference and I suspect that it is purely RF field strength effects.  It doesn't happen on any other bands on the tribander, on the 12/17M 2el yagi or either set of full-size multiband dipoles.  

The shack earth is a 1.8M grounding rod with about 300mm left protruding and connected to the angle aluminium bus-bar along the back edge of the operating table using LDF4-50 coax outer about 1.2M long end to end - so it is very direct. RG213 outer braid flyleads connect the computer, monitor, Icom CT17 CI-V adapter case, radios, coax antenna switches etc to the bus-bar so all are bonded together and to the ground. 

My current solution is to turn off CAT control in WSJT-X configuration as soon as it happens the first time simply by setting the model to None.  I know I will have to spend more time solving the issue properly but just haven't got to it as yet.  Maybe a USB-COM adapter (rather than a COM1: /CT17) is the answer - I have a couple of these but haven't yet tried one to see if that resolves it.

Hope your QRM effects all get solved soon.

73 Doug

[VK2CSW]

I did some more chasing this afternoon while the house was otherwise unoccupied (for the sake of household harmony).

The power lead has already had 2 x clip on ferrites and for extra surety, today, I've put third clip on ring type with three turns of the lead through it.

I will need to recheck my tests, although I repeated them a few times, it doesn't make sense to me.

To test I set WSJT-X to call CQ and allowed it to run through either 4 or 6 transmitting cycles. I then ducked into the lounge to monitor the soundbar.

First test was to see if changing bands or antennas made a difference. Tried 40m, 20m and 10m via my 20m 1/2 Wave Dipole. Nothing seen on 10m but given the mismatch with the antenna, I am not sure that proves anything as I have no idea how much RF was going out. Both 20 and 40 caused the issue.

I have a G5RV Jnr strung up inside the roof cavity. This is much closer to the soundbar. Both 20 and 40m showed the gremlin.

Next up, I tried earthing the soundbar. As it is double insulated it was non-trivial, but I ended up grabbing the USB maintenance port shield which buzzed out as connected to the small areas of chassis accessible. It made exactly zero difference.

Now for the weird bit, I thought I would go back to the simplest form I could, as the soundbar has two physical inputs (power and optical signal), I disconnected the optical cable from the sound bar. I set WSJT-X doing it's thing. It ran for two full minutes and nothing.

WEIRD.

Not trusting the result I reconnected the optical cable. RFI was back. Once again I disconnected the cable and it went away.

I reconnected the cable to the soundbar, but this time disconnected it from the TV. At first no RFI but on about the third cycle there it was.

At this point I would normally call it that the problem is coming in via the optical cable, except I really don't understand how it is possible.

So my new to do list is:
1. Redo the tests (I simply don't trust the results);
2. Replace the optical cable with a different one and retest;
3. Try linking the soundbar via HDMI and retest (this seems illogical to cure RFI); and
4. If all that fails take up Lawn Bowls because this is making my head hurt

Is there any point in putting ferrites on the optical cable?

[VK4ADC]

Colin

When you re-run the tests, try wrapping the sound bar in alfoil (Faraday shield style) with the optical cable connected as normal. It might still be direct radiation into the audio circuits in the soundbar and the shielding might help prove it.

Otherwise, quite rightly : WWWEEEEEEIIIIIIRRRRRRRRRRRDDDDDDDDDDD

Doug

[VK2CSW]

Hi Doug,

Yes that will be a fair test - I have no faith that the test with the optical cable will be repeatable.

Colin

[VK2CSW]

I decided to take a slightly different tack.

I dug out and reassembled a HVU-8 antenna which has languished in the shed for a few years. Well I assembled most of it, the 3.5 MHz radiator did not survive the storage.

I mounted it on a stand in the backyard and made a couple of FT-8 contacts, no evidence of RFI in the sound bar.

I then took it up onto the roof and mounted it. Ran the coax and again tested for interference. All seemed clear.

Unfortunately, SWMBO came home early from work, so any further testing (especially retesting with the dipole) was curtailed.

However, I may have a solution to allow me to get on the digital bandwagon and maintain marital bliss.