[ present & archived diaries ] [ Home ]
I also cut down some of the taller young trees (after the photos below).
I made some forward and reflected power readings at selected (mainly Heating) frequencies before and after the work. There seems to be a slight improvement in VSWR, and in output power at some frequencies, although I would not like to put too much weight on these since they are simply read from the meters and I do this so rarely that I dont know how consistent such readings are between measurements.
As a result of this I have tentatively changed the Tx attenuation, TxHat from 8 to 7 dB. The sniffed RF output seen on an oscilloscope still shows some slight deviation from a sine wave at some frequencies though.
I used the opportunity to make some photos which are now available at http://www.eiscat.uit.no/dynasond.html and click on Pictures on the last line, which is just a link to http://www.eiscat.uit.no/heating/Dynasonde/pictures.html
Thinking that it would be an Ailtech synthesizer problem,
I held off investigations until I had time.
Today, with Assar Westmann (EISCAT HQ), I did some tests to
localise the fault. Using rxbrowse and a test set of frequencies,
it was clear that when the frequency should have been:
2000, 2010 2020, 2030, 2040, 2050, 2060, 2070, 2080, 2090 kHz:
the frequency LEDs showed the following :
2010 2010 2030 2030 2050 2050 2070 2070 2090 2090
It seemed that there was a bit hanging in the tens of kHz
outputs of the particular circuitry driving
the Ailtech synthesizer and the digital frequency meter.
The particular unit, the Control/Adder module, at the
right of the bottom NIM bin with the frequency display LEDs,
was removed and inspected. I dont think we have ever opened
this module before. We immediately found some of the wires
going into the PCBs not soldered properly on
one side of the PCB, while looking OK on the bottom (2 wires
at least but we re-soldered about 4). These PCB boards are
not plated through the holes. Soldering the wires and re-installing
the unit caused the problem to disappear.
The ionograms looked better and the synth status word was 7
instead of 239. Racktest gave zero errors.
In fact the synthesizer has shown a status of 239 for a very
long time, and has always alternated between 239 and 7,
which is why I asked Mike Rose a year or two ago what should the
status be and he said it should be 7. I never followed up
why we were getting 239 often.
The fault, perhaps intermittent, must have been there for
very many years, perhaps since the beginning of this sounder
which was the prototype I believe. I wonder how many soundings in
the past have been affected.
In principle we could re-analyse the affected data correctly by
telling the software the correct frequencies. However it may not
give good solutions with these faulty frequencies.