Tuesday, 25 January 2011

Softrock Pre-Amplifier/Buffer

The short of it is the antenna is 0.5m long but using it with a remote Preamp gives a 0.2-60 MHz receive only  bandwidth.  Version 3 at the bottom is the final build.

Veecad is brilliant for strip board production as is Tinycad for schematic layouts.  I used both of these for the first time to produce the circuit board layout on strip board.  I won't know if RF will like this layout but I will try anyway.  I have no links just components and have kept everything tight.  The schematic and Veecad layout is now complete. Circuit is now built ready for testing.
         My Strip board Version 1

The Veecad Pcb layout below is Version 2 as I got the pins of the FET transistor crossed and had to twist the transistor to get it in.  Also covers new Q1.
 Version 2   PDF Build Instrs

                 Fun with Copper Tape for Screening

Today I used the Advantest R3361C Spectrum Analyser 9 kHz-2.6 GHz used in the Electronics department, of which I am not a member of, at work.  The output of the amplifier was as follows: @1 MHz -24 dBm,
5 MHz -12 dB, 9 MHz -10.5 dB, 13 MHz -13.5 dB.
The results were made into 50 ohms and show more loss at low frequencies.  Using an oscilloscope I found the problem.  I had inadvertently swapped the input and output capacitors.  A 0.1 uF capacitor  was soldered across the output capacitor and the measurements were repeated as follows: @1 MHz -5 dB, 5 MHz -6 dB, 10 MHz -7.4 dB, 15 MHz - 9 dB, 20 MHz -10.5, 30 MHz -13 dB, 50 MHz -16 dB, 100 MHz -20 dB.
I have now swapped the offending capacitors.
So, at 1 MHz I have increased the gain by about 20 dBm (good fix).  I think the Softrock input is not 50 ohms so the gain will be better than I have measured here.  Now to try it...

Initial measurements were made across 80m, 40, & 20m by injecting from my signal generator into the softrock directly.  I then repeated the measurements with the preamp in circuit.  I used the power meter readings from SDR software to compare the signals.  The readings varied no more than 1 dB when comparing with and without preamp.
I made some WAV IQ recordings to try and show the difference with the preamp in and out of circuit.  The band conditions make this difficult but an AM station definetly seemed stronger with the preamp in.

I'm sure the sensitivity of the receiver has been improved and I shall make more measurements over time to prove it.  I think enclosure screening is definetly required before further measurements can be made (now done).

24-11-2010.  Further tests today with a scope.  I put a 20 dB attenuator into the input of the preamp and fed it with my miniVNA in frequency generator mode.  The output was attached to a 330 ohm resistor and the scope was attached to the input and output accordingly.  I also had a look at the output of the NFET and found the gain dropped with frequency so its the first amplifier stage output is not flat.  Frequencies <5 MHz are amplified and >5 MHz are attenuated.

29-11-201.  Second stage transistor was replaced with a BF199 with fantastic results 1-85 MHz flat response with 2 dB gain.  I was extremely happy with this result for such a simple change.

Credit goes to Nigel (electronic boffin) at work, he said 'The FET should have reasonable wide bandwidth but is a high resistance part. It is probably being loaded by the input capacitance of the following stage which due to miller capacitance can be quite large.'

The miniVNA images that follow had a 20 db attenuator feeding into the preamp.  Now perfect for the Softrock!  The first two miniVNA images below show a vast improvement.
only 0-5 MHz with PN2222
0-30 MHz with BF199
0-85 MHz with BF199

Rocky seems good for measuring performance and the Peak was flat all through the HF HAM bands at -36dBFS with the preamp fed with a -20 dB Attenuator.

0 dBFS indicates the digital number with all digits ="1", the highest possible sample.
The lowest possible sample is (for instance for 16 bit audio):
0000 0000 0000 0001, which equals -96 dBFS.  Therefore the dynamic range for 16-bit systems is 96 dB.
Note: the noise floor is at -94 dBFS.

1/12/2010 Copper tape was used to screen the circuit see photo above.  The coax input was moved nearer to the Pcb.  Both input and output connectors are directly connected to the Pcb.

7/12/2010 The time signal at 9996 MHz and the beacon tone at 10 MHz was monitored.  The signal-to-noise ratio was about 2-3 db better with the Preamp in circuit.  The signals monitored were only just out of the noise.  I used the Peak hold facility in SDR software and measured with the cursor facilty.

14-1-2011 Version 3

The telescopic antenna is 535mm long when fully extended.
This final preamp version is for remote operation from my T-Bias power supply.

12V T-Bias with preamp ready for test.
I measured the performance on my miniVNA and got gain from 1-65 MHz.
On 80m & 40m about 5.5 dB gain.  Good down to about 200 kHz and up to the 6m band.

25-1-2011 Isolation of Antenna from Softrock.
Today I measured the SDR Local Oscillator which finds its way back to the input of the Softrock and hence to the Antenna.  At 7.2 MHz I measured 4.83 mV or -33 dBm (50 ohms) and at 3.7 MHz 3.45 mV or -36 dBm.  This was measured with a Yokogawa DL1540C Oscilloscope connected to the input of the Softrock.  I then added the version 2 Preamp between the Scope and the Softrock input and remeasured.  At 7.2 MHz I got 700 uV or -50 dBm and at 3.7 MHz 149 uV or -64 dBm.  I don't think my measuring technique was quite correct but it definetly shows improved isolation of the Antenna from the local oscillator by some 30 dBm.

No comments:

Post a Comment