U3S and 1-80MHz 5W eBay

Last year I have build 3 QRP-LAB U3S multimode beacon transmitters.
The goal was 5W on all HF and 6m to be used as WSPR beacon on a remote location in PY
This goal was never reached. the QRP-LABS 5W PA could only make 1W at 50MHz
3W at 28MHz it delivers 5W at the low bands only.
The QRP LAB "5W PA" uses the IRF510 or RD15HVF1 with 9:1 wide band transformer out
As simple as can be. The RD15HVF1 is a VHF 10W FET it should easy deliver 5W at all bands up to 144MHz
f.i. the excellent 144MHz XVTR (@ transverter-store) uses the RD15HVF Its output is over 10W But whatever i tried with the QRPLAB 5W PA at 12V 3W at 50MHz was the max
I need a conservative reliable 5W that can run 24*7
I gave up the time consuming experiments had to accept the 3W at 28 MHz 12V

Recent found some eBay RF amplifiers that looked promising.
The 2W 1-930MHz f.i. 2 of them could make 4W?
But 0.5W maximum carrier at 5V not the stated 2W
useless as PA for a U3S

Another 2 -80MHz 5W amplifier would be the perfect one if it does what it should.
More expensive 50$ mechanical and SMD soldering looks sloppy.

But it delivers 5W almost over the whole range 2-80MHz even more, 7W at 28 and below
Still reasonable output 2W? at 475kHz , my power meter is not specified below 50MHz
Supply 11.5V 1A 1mW drive
The bias is high 0.9A it can be adjusted lower but then the output drops to 3W
The RF and PCB design is well done that is what counts most.

Device stamps are removed Source is pin2/flange it can be nothing else but a Mitsubishi RD15HVF1
The Chinese are wise enough to use a REAL RD15HVF1 for their products.

The U3S output had to be reduced to almost zero for the maximum allowed 1mW
I reduced it with the bias of the BS170 that is a very sensitive tuning and also not correct for linearity but since the RF signal is a TTL output the following amplifier can not make it much worse..

By mistake the drive was up to 50mW the preamp HEMT xxx was blown
The regulated 8V also used for the RD15 bias was shoted
The power FET was still okay but bias was down 100mA no output.
I removed the preamp guess it's a SPF8951
Made a simple 3winding coupling loop from input SMA to the input transformer and turned the U3S drive up to about 50mW . Bingo 5W and more
A single RD15HVF1 driven by the U3S can do 50MHz 5W at 12V
TX 12V is switched with P-ch FET U3S
I also switch the tx 5V for the onboard BS170
To have power at TX time only high bias is no issue then the PA stays cool at 5W RF

Why can the QRP-LABS RD15HVF1 not deliver 28MHz 5W ?
This Chinese PA can do 5W easily at 28 50 MHz and 3W at 70MHz
But the reason must be the difference of the broadband transformers
This one uses 5 winding's untwisted double wire in and out impedance matching.
The QRPLAB 5W has only the output 9:1 transformer
This is a twisted three thin wires 15 turns in a torriod
Works okay for the low bands 160m 80m 5W+
but reduced output at 24.8 28 50MHz and up

This P-channel switches 12V supply at TX time only, reducing power loss heath and QRM