On Tuesday, 6 August 2024 15:17:16 EEST you wrote:
Hallo Adrian,
DMR Tier III is mainly running smoothly in the software side here. Now I
want to make the next step.
Hello Christian,
Answers below:
I was able to aquire two narrow bandpass filters for input and output. A
little too much loss for production use, but TX/RX isolation is
definitely good enough for testing.
For my prototype SDR repeater based on the LimeNet Micro v1, I have a tunable
6 cavity UHF Procom duplexer. These guys can be factory / home tuned for
multi-carrier operation if all carriers fit within 200 - 300 kHz. The downside
is a little increase in insertion loss both sides, but they provide a pretty
flat response over 200 kHz and excellent isolation. Used duplex split here is 8
MHz so beware there are several variants at the Danish producer.
You'll also want a good bandpass filter on RX in addition to the duplexer. The
SDR front-end is too wide-band and will pick up a lot of out of band signals
otherwise de-sensing the receiver. I can't tell you whether to use an external
LNA in addition to the internal one, because it all depends on some
measurements you'll need to take yourself as specs can vary a lot. I do
recommend a signal limiter before the internal LNA as nearby signals will
overwhelm the RX and create IMD in the LNA.
Also, a friend gave me a suitable 70cm preamp that is waiting for testing.
Do you have an opinion about the best PA?
I'd recommend two amplification stages, and I can tell you what I use as a first
stage: DK6JL's design, to bring the -3 dBm output of the SDR towards 25 - 26
dBm total output (depending on the number of carriers I end up with 17 - 21
dBm per carrier.
For the second power stage, I'm not really there yet due to lack of time, but
I'd recommend to keep it a little in the linear-ish regime (if using a class
C, at least 6 dB back-off), or even a small linear amplifier. I recommend
testing the output on a spectrum analyser for any IMD, as the multi-carrier
waveform is very finnicky. You can safely use a Kuhne linear amp, they can take
as input power around 0 dBm or up to 10-20 dBm AFAIK. To reduce the price, a
second stage using a class C design is desirable though.
My first stage PA source website;
kleinleistungsverstaerker-fuer-adalm-pluto-hackrf-lime-langstone-usw/
Currently, with 100% TX gain in qradiolink, I roughly get -19dBm per
Carrier directly on the TX connector of the Lime
Strong recommendation to not exceed 94 on output gain in qradiolink, the
LimeSDR PA tends to introduce distortion above about 96 gain at least on my
model. Do check the waveform and spurious on a spectrum analyser at that
point, highly recommended.
With a small chinese amplifier (WYDZ-PA-1M-750MHz-3W) I can amplify
this to about +15dBm per Carrier.
Like I said, just make sure to check the output for spurious and IMD, because
of the specifics of the waveform. Cheap Chinese PAs tend to have widely wrong
P1dB figures quoted in marketing, so be careful, I got burned on a couple of
such Ebay purchases which distorted very early, but others were fine if I
backed them off 3-4 dB below their quoted P1dB.
I did not check the signal quality yet in depth with multiple carriers
transmissing at the same time.
From your experience:
- What would be a reasonable TX gain setting for the Lime Mini 2 to get
7 clear carriers out of the SDR?
Maximum 94 tx_gain on my Lime v1. Can't say for sure on the v2 as I don't have
that model.
- What amplifier chain would you recommend to get the output signal to a
resonable level?
See above, two stage amplification, or even one stage using a Kuhne linear
amplifier. You can experiment with a class C as a second stage (FM amp) and let
us know if results are acceptable when backing off the max output by at least 6
dB. I was going to do this this spring but was too busy. My expectations are
that a normal Mitsubishi PA module with adequate cooling should work...
In the beginning, 10-100mW per Carrier would be fine. At the end I am
aiming for at least 50W for all 7 carriers together, so at least 7W per
carrier, maybe more to account for filter and cable losses.
Yeah that sounds reasonable, that's what I'm aiming at too. You're one of the
pioneers here, not many people even attempted this yet, so would be great if
you document your findings during the process and publish them.
On the testing method: What's the easiest way to get qradiolink to
produce an not necessary modulated FM signal all the 7 channels
FM carrier not possible AFAIK unless you go change the GNU Radio code inside
it. But you can set in MMDVM beacons to run for 29 seconds every 30 seconds
and you'll achieve the same result. See my video here:
Thanks in advance
Christian DB9CR
Cheers and 73s
Adrian
ADILSON MATOS