Unsteady Journal of Jolty Notions

Some while ago I was visiting back a former workplace of mine. My colleagues there had just heard about the RTL-SDR stuff and were testing a DVB-t USB stick they ordered from a chinese dealer which was delivered on that day. Since they had a litte bit difficulties in getting all the drivers patched up, I offered my help because I already gone through that experience.

While doing that, I just briefly saw that someone modified the Stick to get LF to HF reception. As soon as I got home I ordered a new dvb-t stick to replace the old one and as soon as it arrived I modified the old one. I first got some shortwave stations but had severe problems with overloading from nearby 88-108MHz FM Broadcast stations. After all I live in the middle of a nations capital, so there is no possible way that this could not happen. I mean the main source of those signals is the Berlin TV-Tower, a mere 3.4 km (2.1mi) away with many stations in the Kilowatts region.

I first suspected the still on board Tuner IC to be the bad part, still making some “conversions” and plunging in the radio stations. But that was not the case, I found out after removing it completely, leaving me a USB stick with only a Realtek 2832u. I still had all the Shortwave spectrum covered with FM radio.

[I really need to get a picture of the spectrum with this, Shortwave and FM side by side 🙂 ]

First I was shocked and did not understand why, later I tried to investigate the reasons and discovered that the stations appered every 28,8MHz in their regular order and every 14,4MHz in reverse. Looking a little bit into this I soon realised that the board has a 28,8MHz crystal oscillator on board, a pretty standard frequency for devices connected to usb.

I once briefly tried to disconnect any antenna and all the FM stations just dissapeared. So this was not some spurious interference coming from my USB-cable extention, it really came via the antenna. No wonder, the antenna is directly connected to the receiver. No preselection is made at all.

In the mean time I even got a Funcube Dongle Pro Plus and although it a really good receiver for most of the lower frequencies, it is pretty deaf. I.e. I can not hear the local 177 kHz Longwave station at all, whereas the modded DVB-t dongle does pretty good  job at it.
But as said, it gets disturbed by the higher frequencies, so :
I just needed to fix this! 

I went ahead and fired up RFsim99 and Eagle Layout and a couple of hours later this was the result that landed on instagram :

Transition frequency of S21 starts at 30 MHz. Since the dongle cannot recieve anywhere above 28MHz there shoud be no problem with the 10m band anywhere. -3 dB is at 33.5 MHz and -40 dB is at 45 MHz. By the time I reach 87.5 MHz, which is the lowest end of FM broadcast here in Germany I have -124 dB, which was particularly important to me. A filter I had built two years ago could reduce the FM interference but not block it completely. I guess I had made the filter slope too flat. Also I was trying to squeeze that between the FM band and the airband, not much space in there.

But just what this should do according to theory is not enough for me. I really want to also measure the S21 and S11 values so I can compare them to what I designed for. Any ideas for that ? What can I use ?

I also sent in the layout to get some boards, they arrived yesterday and wait for me to populate them. So there will be some updates to this. Until I can post anything new on this, here is a photo of the finished pcb. That is some awesome quality for just 29,90 €.

Looking at the PCB I see I labeled it with 60 Mc. Why did I do that ? I can remeber writing Mc instead of MHz because I needed the font to pass the Design-Rule-Check, therefore I needed it to be bigger, until I later discovered that I did a mistake. But I do not recall why I wrote 60 that is neighter the beginning nor the end of the transition frequency. Just in case you wonder, at 60 MHz this should be at -75 dB with tolerances between -68,8 and -81,6. I really am curious if that will come out as expected.