Hello everyone. Though some have been fooled with our brief February thaw, veteran Wisconsinites know that we "ain't out of the woods yet". As we have our fair share of the cold and snowy left, it is much easier to think about the fact that spring is "just around the corner". As much as the winter is a bother to some of us, it has yielded some good information.
I received a brief note from Denny, K9AEG in Wittenburg a few weeks back and what he had to say was most encouraging. Some may remember that the node stack in Wittenburg is an outdoor system, and does not have the benefit of being housed in a nice, temperature-stable environment. It does feature Emergency Power, and an in-cabinet heater should it get too cold.
Denny tells me that recently he had a bit of a problem with the cabinet heater, unbeknownst to him. Apparently, the radios and TNC's were stable well down to the zero-degree range! This is a bonus for us, as we were wondering just how cold a system like this could get before it stopped operating. Now we know that Zero is about the limit.
I told you last month that the ice buildup on the antennas at that site rendered it only partially functional. As expected, we had a warming trend that thawed things out nicely. The Wittenburg stack did in fact return to it' s old self shortly afterward. Of course this prevented a more extreme check of cold-weather operating effectiveness, but it also allowed Denny to find his cabinet heater problem.
So the current favorite combination for backbone nodes seems to be even better than we thought. A lot of Node operators are using this combination of the Motorola Mitrek commercial surplus radio with a Paccomm Spirit II TNC. So far there have been remarkably few failures. Once the crystals age, they seem to be very stable. At 9600 baud, they have to be. Frequency error at that data rate will surely cause significant data decode errors. Luckily, we have been fortunate because we have chosen good equipment.
Speaking of data decode errors, I would like to share something with the folks who are trying to improve the performance of their packet radio systems. I have been searching for the cause of data decode errors in my own station with mixed results, but do believe that the answer may be just around the corner.
The decode errors can be seen if one turns the "passall" for the TNC on, allowing packets that do not meet their checksum to be displayed on the monitor screen. One can see mangled call-signs, funny SSID's, and other interesting packets when this is done.
I have changed radios, TNC's, computers, and serial data rates all in an effort to determine just why this is happening. One change that did produce a noticeable reduction of data decode errors was the changing to a 3-wire serial cable. This, on the surface, would not seem logical. After all, a full serial cable would provide "handshaking" signals that would decrease the chance of bad data exchange between the computer and the TNC. But there is one addition that is not welcome in the audio world (remember that a TNC decodes a change in audio tones!). This is the addition of another ground path in the cable.
Folks who have had experience with audio signals have been exposed to the "ground loop" problem, no doubt. This is the presence of hum and noise caused by two differing ground paths between audio equipment. After all of the investigating that I have done, it appears that this indeed may be causing some of the data decode problems that I have been experiencing. Let's look at how a typical Amateur Radio station is connected, and see how this relates to ground loops.
First, we all like to ground our equipment. Lightning protection and the presence of strong R.F. currents are the primary reasons for this. But consider how most of our packet radio stations are wired. First of all, the computer is often grounded not only through the power line, but maybe through an additional shack ground. The serial cable has at least one ground wire connecting the computer to the TNC. The TNC is grounded, through the power supply and sometimes additionally through the shack ground. Then the radio is connected to ground, generally through the antenna system and shack ground.
Complicating matters further is the fact that may hams (myself included) use the same power supply for the TNC and the radio. The negative lead of the power supply connects to both the chassis of the radio and the chassis of the TNC. So now we have 2 grounds connected between the radio and the TNC in addition to the wiring connecting the TNC and radio together. This is where I think the problem lies, and where an improvement can be made. All of these ground paths inject hum and noise into the audio signal, causing data decode errors.
The next step is to couple the TNC and Radio together without ground loops. I have procured transformers and a relay to do this with my station, and will be trying this soon. If you have encountered this problem and have not solved it, or have some experience with this, stay tuned. I will relay the results of this experiment next month. Even the most elegant of theories has to be proven scientifically.
The last thing to remind you of is the quarterly WAPR meeting. There is no firm date and place as of this writing, so do keep checking your local BBS for this information. As always, anyone with an interest in Amateur Packet Radio is welcome to attend, whether they are a member or not.
That is all I have for this month. As always, I invite your comments, suggestions, and input. My BBS address is: KB9ALN@KB9ALN.#GRB,Wl.USA.NOAM
Until next time, 73 from Andy