mandag den 18. maj 2020

Noisy PC Speakers - and a Solution

My PC-speakers are pretty standard: A pair of Creative A50 powered with 5 V DC from an USB-port.

I noticed noise coming from the speakers, when no audio signal was present. Changing USB-port did not help. I seemed like my USB ports were noise polluted! When I connected a power-bank (lithium-ion battery) to the speakers, the noise disappeared, but during pauses in the audio, the power-bank shut itself down!

I decided to build a noisefree 5 V DC PSU using a transformer with iron core and a 7805 which has 68 dB ripple rejection and low output noise. The 7805 is thermal overload protected and short-circuit protected. I had all parts laying around, so the building cost was nil.

Figur 1. The noisefree PSU.

Figure 2. Noisefree PSU with open lid.

Figure 3. AC Adaptor. 
Figure 3 shows the transformer which is an old AC adapter with iron core. It is heavier than modern AC adapters! The advantage: no switching technology is involved!

Figure 4. Circuit diagram.
The rectifier bridge is an integrated piece, but four discrete 1 A diodes can be used as well. I found C1, C2, and 7805 in the drawer. The circuit board has holes and solder islands, but no tracks. Figure 5 shows how the component wires are connected on the bottom side. The alu-cabinet is Hammond 1590A.

Figure 5. The circuit board seen from the bottom.
The PC speaker's power consumption is low. The noisefree PSU does not get warm at all.

Now I can enjoy music and speech without background noise - and a good feeling of having improved something!

73 from OZ1BXM
My homepage: http://oz1bxm.dk


 

tirsdag den 24. marts 2020

Replacing Meinberg NTP with Network Time

I purchased a new PC last month with Windows 10 preinstalled. The Windows specs are shown in figure 1.


Figure 1. My new Windows 10.

I had problems after I installed Meinberg NTP. Whenever I booted the PC, the system clock was set by Meinberg, but then time discipline was lost. After some hours, the system clock could deviate as much as 1 sec. It seemed like Meinberg could not control the system time. I worked with this problem for a while, and then I decided to uninstall Meinberg NTP and try another piece of software for time control.

I need accurate time keeping because I'm using weak signal digital protocols in amateur radio. A time deviation of less than 100 ms is required with JT65. I use JT65 when I'm bouncing radio signals off the Moon and back to Earth (distance 800,000 km).


Figure 2. Network Time.

Network Time was easy to install, and it worked right away. Under "Settings" I set "Update Interval" to 20 minutes because I wanted frequent updates. Remaining options were left at the factory settings.

Before installing Network Time I opened UDP port 123 (inbound/outbound) in Windows Defender Firewall. This port is used by the NTP (Network Time Protocol). I named the rule "UDP-port-123-time1", see figure 3.



Figure 3 Windows Defender Firewall (Windows 10).

I changed "Set the time automatically" from On to Off as shown in figure 4. This tells Windows not to use W32time.


Figure 4. Date and Time (Windows 10).

You can check the accuracy of your computer clock by visiting this page.

Vy 73 from OZ1BXM Lars
Webpage: oz1bxm.dk

onsdag den 12. februar 2020

2 GHz Spectrum Analyzer with SDRplay

SDRplay RSP1A can be used as a spectrum analyzer between 1 kHz and 2 GHz. "RSP Spectrum Analyser" is free Windows software developed for this purpose. "RSP Spectrum Analyser" can be downloaded from the SDRplay website.

Figure 1. SDRplay RSP1A with a 20 dB attenuator.
I wanted to test a Pierce crystal oscillator. The circuit diagram is shown in figure 2. I've built the oscillator on a piece of Veroboard using leaded components. The crystal frequency is 7.030 MHz.

Figure 2. Pierce crystal oscillator.

Figure 3. Measuring the crystal oscillator.

The different items are connected as shown in figure 3. A is the crystal oscillator. B is an active probe; it has no amplification but Zin is 10 Mohm and Cin is 0.5 pF. C is a 50 ohm coax cable. D is an 20 dB attenuator. E is SDRplay, and F is the USB-cable connecting SDRplay to a Windows PC.
Figure 4. Screen-dump from RSP Spectrum Analyzer.
The screen-dump in figure 4 shows a 5-25 MHz sweep. On the vertical axis, the reference level is 0 dBm and the lowest level is -80 dBm. Spike A is the fundamental frequency, C is the 2nd harmonic at 14 MHz and E is the 3rd harmonic at 21 MHz. The remaining spikes are B, which is a spurious signal, D is a real signal, and F is a spurious signal from the SDRplay clock at 24 MHz. Spike F can be removed be activating "clock spur removal".  

Another example is measuring a band-pass filter for 1090 MHz. I connected a noise generator in front of the filter in order to create an input signal. The RSPplay was connected to the filter's output, and no attenuator was used this time (the noise signal is weak and should not be attenuated). 

Figure 5. Measuring a band-pass filter.
The reference level in figure 5 is -40 dBm, and the lowest level is -120 dBm. The sweep is 1000-1200 MHz. The filter curve is clearly visible. 

I've had much fun running SDRplay as a spectrum analyzer. The RSP Spectrum Analyzer software works excellently. For the radio amateur and the hobbyist, SDRplay is an acceptable alternative to dedicated (and expensive) spectrum analyzers!

73 from OZ1BXM Lars
Homepage: http://oz1bxm.dk/