250 Watt 1GHz resistor

[Jan Panteltje]

On a sunny day (Sat, 04 Oct 2014 10:01:29 -0700) it happened Jeff Liebermann
<jeffl@cruzio.com> wrote in <jv603al0qicp0omco5so5emb9mdhptvn45@4ax.com>:

cooking or melting baluns. That I was ranting about this time is that
putting a 75 watt transmitter, into an undersized antenna with serious
sidelobes, in a metropolitan location, is going to create interference
for other users of the frequencies. Much depends on how much effort
you put into preventing interference. Simply claiming that it won't
happen and there's nobody nearby to complain is insufficient.

As for the balun, most baluns exhibit about 0.5dB loss. The typical
barbeque grill dish antenna has a balun between the dipole elements
and the nearby N connector. It's made for receive, so there's no
problem with an additional 0.5dB loss in the balun. However, for
transmit at 75 watts and a properly matched system, the balun will
dissipate 11% of your power or about 8 watts. 8 watts through a very
small PCB is going to burn the PCB. There's no easy way around the
loss problem if you use a balun, which is why I suggested a waveguide,
horn, cantenna, or patch feed, which eliminate the balun.

You should really not rant so much.. its bad for your health..
I was contemplating this antenna:
http://www.produktinfo.conrad.com/datenblaetter/975000-999999/995897-an-01-en-TP_LINK_PARABOLANTENNE_24DBI_ANT2424B.pdf
Frequency Range 2.4GHz ~ 2.5GHz
Impedance 50
Gain 24dBi
VSWR 1.5
Horizontal Beamwidth
Vertical Beamwidth
14°
10° U
F/B Ratio 30dB
Polarization Vertical or Horizontal TL
Maximum Input Power 100W <----------------------------------------------
Connector N Female 24
Application Outdoor
Mount Style Pole Mount / Wall Mount
Mechanical Specifications
Antenna Dimension 600×1000 mm
Weight 3.5 +/-0.15 KG
Mounting Mast Diameter Ř30~Ř50 mm
Rated Wind Velocity 216 Km/h


I have a frquency versus gain graph of it somehere, its still pretty good at 2.4.

We have a great working coversity network.
where the transmitters are in sync via GPS,
I can reach the whole NW and NE side of the country with my porto.
http://www.coversity.nl/in.html
look a the map.
There is even an android ap that can show which repeater I am on / is active at the time.
Of course it is cheating, as data goes via cable, but it sure unites a lot of hams, it is a busy channel.

Very nicely done. We have various linked repeater systems. I don't
see much use for them, so I don't participate. However, one local
system that has my attention is a voter and simulcast system on
440MHz:
http://www.wb6ece.org

Interesting, seems more or less the same idea.
One thing I noticed is that sometimes you get echo...

Here if I am with the porto on the south side of the house I trigger a repeater on the north side,
and on the north side of the house triggers the repeater in the south,
logic Mr Spock,?? maybe the walls reflect..

What is a problem here is internet over mains wires, if I hold my RF detector next to the mains it goes wild.
You could buy those little ethernet plug in boxes, not sure who has those in the neighborhood.
There goes your short wave.

You might try some kind of low pass filter where the power enters the
house. Maybe a big ferrite bead. It won't do anything for radiation
from the power lines, but it will keep the RF from entering the house.

He have plenty of HomePlug devices in operation. In California, there
are only a few BPL/PLC installations. Locally, the major sources of
HF RF noise are:
Solar inverter installations.
Switching power supplies and chargers.
LED lights.
Computer noise.
I have all the above in my house. When I want to operate HF, I have
to turn off almost the entire house.

Yes same here, I scrapped an ebay 12V adaptor,
the seller showed one with a picture of FCC and CE markings.,
what I received had neither, and even managed to create interference bands in
a screened video link on the otehr side of the house.
no radio reception possible to 20 MHz or so...
So scrapped it, and used the box to build my radiation logger in:
http://panteltje.com/panteltje/pic/gm_pic2/

Was evaluating some uplink designs today, here is one:
(...)
Hope U use a real newsreader... guggle will likely fold it


Forte Agent displayed it nicely, but wrapped it into an unreadable
mess when I tried to quote it for a reply. I could put ">" symbols at
the beginning of each line, but it's too much work right now.

Agent was a good newsreader, but no Linux version,
so I wrote NewsFleX, back in those 199X days:
http://panteltje.com/panteltje/newsflex/index.html
the GUI is inspired on Free Agent, originally it was called Secret Agent ...

Nice transmitter idea. I haven't worked out the filter requirements
to prevent mixer spurs and unwanted harmonics from being transmitted.
Designing a transmitter like that is actually fairly easy. Finding
frequencies that don't create interference is difficult. You'll
probably need some bandpass filters somewhere.

The problem with satcomm (I am NOT an expert on this) is not the
transmitter. It's the receiver. The downlink is on X-band (10GHz)
which will work with TVRO Ku band components but usually requires a
wall of test equipment to get things right. This is typical:
http://www.uhf-satcom.com/xband_hw/

Yes but the downlink is at 10 GHz, so my normal dish with a modifed LNB should do.
I found some interesting things, been studying modifiying standard LNBs,
I have a plan (more than one actually).

Maybe or maybe not, I also ordered some ceramic SAW 400MHz filters, 7.8 MHz wide on ebay, those SAWs cost next to nothing.
This method above needs no steep filters, the U2790 only produces 1 sideband, no carriers,
look up the chip.
The 473 is easily filtered out in the 2.4 GHz filter.
For this I have - or have ordered most parts,
and there is still space on the board, I had the U2790 on board and removed it, it will be back.

This is one of 3 topologies I came up with, and the first one I will test.

Good luck.

I managed to transmit a 45 minute or some mpeg2 movie yesterday, until things stopped working.
Late last night scoping I found the flatcable that connects the raspberry to the modulator board had intermittent
dataline 0 (last pin on cable), plugging and unplugging the connector many times broke the flatcable inside.
I just managed to get that connector of and put it back on again (no spare at hand).
Did you know Chinese flat cable looks thinner (the copper inside) than the normal ones?
I mean I have some USB cables with hair thin strands, but they do it to flatcable too!

Did some VCO stability (drift) tests yesterday, build a second stabilized supply for the VCO only,
These things have a 10 to 20 minutes or so warmup even with their own supply.
But it will be in a PLL, it still works without PLL! Pretty good.

The other thing I want to test if I can build a 90 degrees (old fashined way) audio phaseshift and
that way transmit SSB (driving I and Q with 90 degrees out of of phase audio),
as then I can use their narrow band transponder with a resolution (step) of 256 Hz (prescaler),
and work voice via that sat.
That would be the old 'phasing method'.
Hey I have to 2016 to get it working.
Also ordered a 2 W 2.4 GHz amp from ebay, that will be to drive the big one, and great for local testing.
Hope that works, sort of assuming those WiFi amps can do QAM and must be linear in a way.

 

[Jan Panteltje]

After experimenting for a while, I decided to change teh setup.
This was the old one:

Qatar geostationay satellite

Linear transponder
2400.050-2400.300 MHz Uplink
10489.550-10489.800 MHz Downlink

Wideband digital transponder
2401.5-2409.5 MHz Uplink
10491.0-10499.0 MHz Downlink




Universal LNB:
Rx IF LO
10700 - 11700 950 - 1950 9750
11700 - 12750 1100 - 2150 10600




Wideband:
VCO
2401.5 - 964 = 1437.5
2409.5 - 964 = 1445.5

Narrow band:
VCO
2400.05 - 964 = 1436.05
2400.300 - 964 = 1436.3

From this a Sirenza 1400-1560 VCO will do for the second mixer.

964 + 9750 = 10714
DVB-S tuner
test out
964 1441.5 - 964 = 477.5 < filtered out 3 x 8 V 20 A
in series
| 1441.5 + 964 = 2405.5 24 V 17
A power
I --- 964 only | |
U2790B -------------------- X ------------------------ 2401.5 ----- bandpass --- 1 W driver amp ---- 75W Spectrian Linear RF Amplifier Board -( dish
Q --- SSB RMS42-H 473.5 2400-2410 |
quadrature ring diode |
modulator mixer UBP1505
| | :256
Sirenza VCO | prescaler
VCO Sirenza 9.3752 - 9.41211
951 - 977 VCO |
set to 964 1400 - 1500 |
set to about phase comparator --- loop
filter ---> VCO2
405.5 - 964 = 1441.5 |

|9.3752 - 9.41211
FPGA ------------------
frequency control - user
|

10 MHz
Rubidium
reference

I decided to go directly for 2.4 GHz

I ->-
AD8346 ------>----------------->-- 2 W driver amp ->- 75W Spectrian Linear RF Amplifier Board ->-( dish
Q ->- SSB |
quadrature UBP1505
modulator :256
| prescaler
^ |
| | 9.3752 - 9.41211
Sirenza VCO |
VCO |
2400-2500 <---loop filter <-- phase
comparator
|
^
|
freq out Raspberry ------>----------- 9.3752 - 9.41211


I had problems with the U2790B quadrature modulator,
actuall the second time, now it seemed the chip is defective and has some offset on one set of inputs,
this is the same problem I had with this chip and why I removed it and replaced it with the mixer setup.
But now I have this spectrum analyser:
http://panteltje.com/panteltje/xpsa/index.html
maybe does not go all the way up to 2.4 GHz, but I could test at 1.5 GHGz, I foudn the Analog devices AD8446 works 1000x better than
that U2790B chip (at least better than this one), gives good carrier and sideband surpression, and was already on board.
I tried to drive it with the DVB-S I and Q signals and got a beautiful signal, even without niquist filters.
As the AD8346 goes all the way to 2.5 GHz (unlike the U2790B that is limited to 1 GHz),
I do not need any mixing... Output level is good too...
2 W driver amp came in, 2400 MHz VCO just came in, 75 W power amp passed the DC tests, MOSFETS OK.

So, removed a lot of components:
http://panteltje.com/pub/creating_board_space_IMG_4585.JPG
the big AD9761 is not used for DVB-S here.. but it allows me to make QAM 256 etc..
QAM with 4 dots is now easy, the 2 pots are for test to feed the IQ signal directly into the AD8346 quadrature modulator.
The extra LM317 in probably no longer needed, it is a separate power supply for the VCO.
LM317s are very very stable.
In this picture there is a low frequency VCO, the 2400 MHz one that just came in will be probably mounted somewhere south of the AD9761.

Very simple.
The other adventure to come is to have Raspberry Pi generate the frequency 9.3752 - 9.41211 to compare the prescaler output against.
If I can make that sufficiently stable (its a fractional divider) no FPGA is needed, and then I can set the output frequency from the command line.
Any offset correction can be done in software, will have to test.
Next is 2 niquist filters for the I and Q signals that will for now be fixed at 1500 symbolrate, this
so I can use that only PLL out of the raspi... then some diff amps to drive the AD8346 symmetrially, some real hardware switches to
switch hbeween VCOs, if that proves difficult add an other AD8346, I still have one somewhere, and use a second prescaler and output
connector (already there).

So are the results of a day with the soldering iron, 4 computahs running to control all stuff..
Now that was good old electronics.

One advantage of not mixing is no mirror frequencies, and a very very pure QAM signal.
I am also contemplating makeing 9 0degrees audio pahse shift with RC networks (see ARRL) and use that to send audio SSB as option on that sat
narrow band transponder.


Hope U use a real newsreader... guggle will likely fold it
>