Sound Card Stability Tests Applied to PSK31
After almost 50 years of hamming I discovered the sound card digital modes just a few months ago. Now after spending a few days trying out SSTV and RTTY I settled on PSK31 as the mode I enjoyed most. Recently I was working a station and got a 599 report but the operator also said he was getting some garbled words. We sort of attributed it to the fading characteristics of the path which is known to raise havoc with a narrow band signal like PSK31. However, I didnít notice any problems copying at my end.
As a result of this I decided to check my signal. I wanted to check the IMD but in the process discovered that a second system I was setting up had significant wiggle in the idle tracks as viewed on the waterfall display. It looked like a bad case of DUI!
Two systems were set up since I have two locations that I wanted to have digital mode capability. The first system use a PC (Pent II, 400mHz, Creative Audio sound card) interfaced to a FT-100D. The other system is a Gateway solo 2500 laptop (Pent II, 400 mHz with a built in sound card) interfaced to an IC-756PRO. Both of these systems have mixW 2.03 software installed.
When I first tested this setup I discovered that the laptop had a significant phase stability problem that was proportional to the audio frequency offset. At 500 Hz the copy was R5 or 100% on the MixW tuning indicator however, at 2700 Hz the signal on the indicator was 199 and copy was beginning to get garbled. At this audio offset the signal, as received by the PC system showed about 45į of phase deviation at idle when viewed on the tuning indicator. Also, the idle tracks on the water fall display had a noticeable wiggle. This frequency or phase flutter was at about a 1 Hz rate.
I decided to do some more tests and especially wanted to check the sound card in the PC. I verified that the IMD was reasonable (<-25db) on both systems. In order to simplify the testing I realized that the problems I was trying to track down didnít involve the transceivers at all. Therefore a direct connection via the audio I/O between the two computers was all that was needed to recreate the original problem: a clean (low IMD), strong (S9) signal that was R1 copy!
The figure below shows the setup between the PC and the Laptop. Isolation transformers were used to minimize ground loop noise between the two computers.
The results of my testing are documented by some screen shots. In summary the Laptop has about a +/- 0.2%, 1.1Hz FM componant in the soundcard generated signal. This degree of instability makes it unusable above about 1 kHz audio offset when considering the added propagation phase noise to the signal. The PC has usable performance but would be nice to have it better since a radio with a broader if filter could provide up to say 5 kHz of tuning range.