Switched LNA

[VK3XDK]

Thanks Claudio,
When i designed the PCB i added pads for the stabilization components (if needed) and i am glad i did!
initial stability was poor without these components but they do add to the NF.

so what i have found now is that "loading the output with a good load" also has similar effect (stabilizing the devices) but does not have the same negative effect on NF.
i have not seen any stability issues since.

very new design, there could be some "hidden bugs" but in general looking good.

certainly appreciate the feedback!!

[VK3XDK]

The devices also run very warm (i wonder if this is causing a lot of the problems PGA103)
i have added small copper "shim" heatsinks to mine. hope to post some photos tomorrow.

[VK5ACY]

I have built one of these preamps for 2m but have had to change (also !) to a PGA103 because I have no tiny hi-Q trimmers for the input circuit.

My PGA103 is running in a 'wide open' state and essentially the VK3XDK PCB is a 'test bed' - I can operate the relays/PGA103 and listen to the
change in noise quieting on a weak FM signal. I have no local transmitters except cellphone and etc., so front-end thrashing is not a concern.

My PGA103 runs warm = 80mA at 5 VDC is a fair bit of dissipation for such a tiny component, but the device tab is soldered to a pad with a via
to shift some of this heat.

When this broadband preamp is placed in front of any commercial rig here a noise quieting effect is quite noticeable anywhere from 28 to 439MHz.

This preamp is currently running as a TV preamp in a 75 Ohm system, most of the time I can now see all of Adelaide TV DVB-T services, whereas
these signals were often pixellated/dropping out altogether....(the preamp will work again as an EME preamp when I score some tiny trimmers.)

Thanks Graham for another good PCB - and I will remind others that the PCB-mount relays are those used in the DSE Kit 100W PAs etc., so there
should/will be many of these TOHTSU CX-120P relays in junk boxes around the place.

EDIT: I had a read of the PGA103 stabilising Application Note mentioned above, but I have seen/measured no apparent instabilities in my setup..
..just lucky I guess ! My preamp is at the 'shack end' of the feeders on both ham gear and TV systems and would have seen some odd impedances
in and out...

What I did find however, is (eBay, cheap) 78L05 regulators died at 40-60mA with a head voltage of just 8V...beware the cheapies...a "1A L7805" used now.

[VK3XDK]

Sorry its taken a while to get some photos up!
Here is the PGA103 version (v2) of the "switched LNA"
gain is around 22db (there is 40db of attenuation on the input for the photo)
I have used thicker PCB with this version (hence the wider tracks), and note the "copper shim heatsinks" (quick, easy and can be applied to other boards/ projects))

second photo shows the tight RX bandpass (for those that dont mind a "bit of skirt" )

[VK3XDK]

I re-ordered the V1 PCB's, they are now in and available. (please send email if interested) grumss@yahoo.com.au

The V2 boards are also available


Price (PCB only) is $6 each. (add postage)

[VK3MAT]

as a beginner here in the rf design world how do you check the stability of your lna?

Cheers Matt,

[VK3XDK]

It's not always easy to detect stability issues,
One of my quick tests is- Under bad conditions (no load on the input) wave a piece of aluminum over (and close) over the components, watching for any signs of unusual spurs (or noise) on the spectrum analyzer.
You can also try the same test in a aluminum box (as we all know- LNA's are notorious for being fine until they are in an enclosure )

Without a "speccy" all that can be done is to tune over a band listening for anything unusual (and see if it disappears when the LNA is removed (ive fallen for "Band noise"/birdies a couple of times so make sure there really is an issue))
Of course at the end of the day, most LNA's that are troublesome will probably be oscillating in an entirely different place. (out of band (depending on the LNA, around 5GHz is a common oscillation area)

To fix a problem is another story, LNA's like good input and output loads- sometimes a bit of attenuation on the LNA output give its "bum" a better match and is worth trying.

Input load instabilities can be much trickier to fix- any mods to LNA inputs are sure to increase noise figure.

When LNA's are designed, the stability is usually checked for UN-conditional OR conditional over a range of loads and frequencies BUT that is with the correct tuning. If an LNA is "misstuned" ( usually for max gain (as many do))- all of the stability calculations may as well be thrown out the window.

Hope ive helped!

[EA4ADJ]

Hi, I'm finishing an pre-amplifier for 2m band.
For now I don´t put any filter at the input of the preamplifier, and at the output I put a bandpass filter, details and photos at:

http://ea4adj.jimdo.com/previo-de-recep ... a-144-mhz/

The box is made of one side PCB, now I'm finishing it.
In a few days I'll post pictures of the finished amplifier.
I'll make a version in tin box for sale.
I liked the result of the tests, I've been comparing with MHP 145 SSB with similar results, a little bit more gain for my home-made pre-amplifier made with pga103 +.
Weak signals that were not seen in the waterfall of "Spectran" they looked sharp with my home-made pre-amplifier made with pga103 +.

And now I'll start another version for 432 MHz
73.

[GM3SEK]

"Here's one I made earlier" - a 2m masthead preamp using low-cost PCB relays which are good for at least 750W:



The idea of using compact AC power relays at VHF was originally developed by G4DGU (of muTek fame) and more recently by HA8ET <http://www.ha8et.hu/Mast_Mounted_EXTRA_ ... ounted.htm>. The AC relays handle the RF power and the small SMD relays add extra isolation. Similar relays are widely used at HF with good reliability up to at least 1.5kW. Extending this performance to VHF is all about the choice of relays and a good board layout.

The rest of the design is based on the RF stages of the Anglian 2m transverter by G4DDK and friends, described in DUBUS last year. Features include very tight filtering, NF typically 1.2-1.3dB (for the whole package including the relays) and gain of 11-18dB (adjustable for optimum dynamic range).

RSGB members will be able to download full details in about one week's time from the new online technical magazine called RadCom Plus.

73 from Ian GM3SEK

[VK3XDK]

Interesting, nice work!
I have been using similar relays for power amp by-passing, they work very well up to and above 2M.
I assume that with your 2 relay system that your input relay contacts are in series, how much does this add to your noise figure?

[VK3XDK]

Here they are,
This is a board that i have been using for some time, it is a PA RX by-passing board (with soft switching) designed to be front panel mounted.
Easily handles 300W on 2M (haven't tried any higher)
(ive been meaning to make them available)

[GM3SEK]

Interesting, nice work!
I have been using similar relays for power amp by-passing, they work very well up to and above 2M.
I assume that with your 2 relay system that your input relay contacts are in series, how much does this add to your noise figure?

The short answer is: "Too small to measure". I'm not trying to be evasive but we're dealing here with a small fraction of a dB, which is embedded inside other small losses, and I don't have a VNA with that level of accuracy at 144MHz. However, we can try to unpick it a little...

The dissipative loss in the "DC/AC power" relay (Schrack/TE RTC31C012) is very small indeed at 144MHz - much less than 0.1dB. We can say that with confidence because there are no signs of overheating in that same relay when transmitting at 1kW.

The small SMD relay is needed to give extra protection to the preamp device, because the DC/AC relay on its own has rather poor isolation to the open contact. This adds another small fraction of a dB in loss, but it still isn't significantly different from many equivalent setups using coax relays.

The measured overall NF - including all relay losses and another ca. 0.1dB loss in the selective input network - is typically 1.2dB which meets the requirements for terrestrial DXing.

The really important thing is that the preamp is designed to make people want to use it at the masthead... because "at the masthead" makes the biggest difference of all.


73 from Ian GM3SEK

[VK3XDK]

Very nice board and very good performance, i try not to be too critical (much of what we are talking is trivial)
the PGA103 device is specified as around 0.5db NF at 144MHz, what is bringing your noise figure up to 1.2? I would have assumed a big part would be input losses.

[GM3SEK]

Very nice board and very good performance, i try not to be too critical (much of what we are talking is trivial)
the PGA103 device is specified as around 0.5db NF at 144MHz, what is bringing your noise figure up to 1.2? I would have assumed a big part would be input losses.

The losses between preamp input and the RF amplifier device are roughly the sum of the losses in the AC/DC power relay (the big orange one) + the SMD isolation relay + the tuned input network (bare yellow coil and NP0 ceramic chip caps underneath the board).

The device is actually an SPF5043, with a claimed NF of about 0.8dB says the datasheet.

But let's be very careful about subtracting the claimed device NF of 0.8dB from the measured preamp NF of 1.2dB. The claimed NF on the datasheet is actually a measurement too, and just about all NF measurements outside of national standards labs are subject to uncertainties of at least +/- 0.2dB.

In other words, when you subtract 0.8(+/- 0.2)dB from 1.2(+/- 0.2)dB, the answer does not necessarily equal 0.4dB!

If we really needed to know all the details of the input losses, we definitely wouldn't be working with NF measurements at all. It would need quite careful work with a VNA... and as I said earlier, I haven't got one that covers 144MHz.

By the way, the article has now been released on the RSGB website (members-only section) and parts will be available from Sam at http://www.g4ddk.com.


73 from Ian GM3SEK

[VK3XDK]

OK, thanks Ian
SPF5043 0.8nF --- that adds up better!