[RTTY] RadioShack isolation transformer replacement

[Don Hill AA5AU]

Got it!

Chen's good, eh?

73, Don AA5AU

-----Original Message-----
From: Kok Chen [mailto:chen at mac.com] 
Sent: Monday, May 14, 2012 3:57 PM
To: RTTY Reflector
Cc: Don Hill AA5AU
Subject: Re: [RTTY] RadioShack isolation transformer replacement

On May 14, 2012, at 1:15 PM, Don Hill AA5AU wrote:

> So is it good practice to connect the radio directly to the soundcard 
> without an isolation transformer if you don't have a ground loop?

"Don't have a ground loop" is probably a relative phrase, Don.

If you have a software spectrum analyzer, you can look at the area around 60 Hz and 120 Hz. 

There should be plenty of free sound card spectrum analyzers for Windows, and Mac users can use cocoaModem's spectrum analyzer or a
standalone spectrum analyzer example in my Audio Library package.

If the hum are down more than 15 dB from your louder received signals, it won't bother your receive direction a single bit, since
the software modem has (or should have :-) a decent DSP roofing filter anyway to pass only the RTTY passband to the actual
demodulator code.  IMHO, the only time ground loops bother the receiving direction is when it is loud enough to affect the blocking
dynamic range of the sound card for regular signals.

So, if you are using FSK, you problem ends there.  I really don't think ground loops are really a problem for FSK users and a ground
strap will certainly bring the ground loop to a manageable level.

However (!) if you are using AFSK (or any of the modern digital modes), even -30 dB hum loops is not good enough.  

At -30 dB, a 1000 watt RTTY signal will have a 1 watt component 60 Hz or 120 Hz from the suppressed carrier (or 50/100 Hz if you
live in those countries).  Lots of people will hear your spurious signal -- which appears as a weak, possibly dirty, carrier about 2
kHz from your RTTY signal if you are using a tone pair around 2000 cycles.  

If you transmit both 60 Hz and 120 Hz components, it will appear on RF as a pair of carriers 60 Hz apart and everyone who have that
pair in their passband will hear a 60 Hz hum.

You can imagine what a -10 dB hum component will do.

-30 dB relative to a 0.7 volt RMS AFSK signal from the computer is not that hard to achieve.  But I have known people who pipe audio
at the 10 mV level into the microphone input of their rigs.  When you do that, hum loops can be larger than the AFSK signal itself
and even ground straps may not be enough and a transformer becomes the necessary evil.

Transformers are "quick fixes."  They are not linear devices by any stretch of the imagination (remember your B-H curves from first
year EE?).  I think it is best to reduce ground loops as much as possible with ground straps, and only then only use transformers in
addition, if you still need it.  And as I mentioned earlier, transformers are probably seldom ever needed for a well grounded FSK
station (unless you need galvanic isolation for other reasons).

Manufacturers like SignaLink and microHAM include transformers probably as a "lowest common denominator" kind of thing to take care
of even the most clueless appliance operator.  At least MicroHAM uses a halfway decent transformer (from IMD viewpoint for wideband
users) in their designs.

73
Chen, W7AY