In the vain of VK3YE’s [[Even simpler Software Defined Radio|]] I built a simple breadboard SDR circuit based on an SA612 mixer.

[img[]] //Photo of the board//

This circuit is based on the SA612, I chose this because I have a few in my parts collection, it provides some gain internally and it has a crystal oscillator circuit built in.

Be warned, this is not special or fancy, it is a direct conversion receiver with no image rejection and no frequency agility. There are no separate I/Q outputs.

[img[]] //Circuit diagram of SDR//

The only additional parts I have put in are capacitors on the output and a high pass filter on the input (else I get to listen to 1323 AM). I decided on the 9V battery as I don’t have any easy way to provide a safe voltage for the SA612.

Crystals I have tried so far: 3.579545, 10.140 & 14.3181. I may also have a look through my other crystals to see if there are any interesting services on HF that might be in range. I also want to give 10.14 another go and change the oscillator capacitor values as the carrier is at 10.1385, which is interfering slightly with the reception of JT65.

!! Receive path analysis |!Item|!dB| |Antenna (40m inv-V, EZnet analysis)|5.83| |Antenna [[mismatch loss|]]|-17.33| |HPF (broadcast block)|<-0.1?| |SA612 (+ 10.14MHz xtal)|+17| |Soundcard (onboard VIA HD)|+30| |SDRadio 0.66 (software)|| |Virtual Cable (software)|-| |WSPR-X (software)|-| ||| |Total Gain|35.4dB|

[img[]] //P-Syn filter analysis//

!! Last night with SSB 80m Performance so far is not really well tested, the best I had was last night after 80m was quiet I could hear a station calling DX on 40m on the second harmonic of the 3.5MHz crystal, but the signal was really really low, enough that I could tune the frequency correctly and calculate the offset from the harmonic and find the signal on 40m, but not enough to make out what he was saying.

!! 30m JT65 I played with the capacitor values, 150pF & 180pF and got the carrier frequency down to 10138.28, and combined with filtering from [[SDRadio software|]] I am using I was able to eliminate the noise by setting the cut off at 400Hz.

!! WSPR See: DeadEasySDRWSPR

!! RX Performance & Adding An LM386 On 30m I was getting a noise floor about -95dB. Adding an LM386 brought it up to about -50dB, which lines up precisely with the 46dB gain of the LM386. This means the LM386 only adds gain and since the sound card input is already very sensitive there is no advantage.

But there is a disadvantage, dynamic range. if a really strong signal saturating the input then the AGC in the SDR will be unable to cope, and the input signal will get clipped, image rejection will become a problem, etc… The only advantage I can see is AM radio signals still seem to be leaking in, and a higher noise floor swamps them out of existence.

So why is my FT817 so much more sensitive? Local noise! Add to that I am receiving signals from both sides of the centre frequency, and the harmonics from the oscillator also add noise from higher frequencies at 20 & 30MHz as the oscillator is still very loud at those frequencies.

!!! 30m Filter I have added a series bandpass filter to the input to see if that improves the noise issue. Filter is just in series with the input, a 39pF cap and 6.8uH choke.

!!! 80m Filter Series: 56pF, 33uH. [img[]] //Super sharp filter for a small part of 80m, should be good for 3.6 - 3.8MHz//

!! Future I am tempted to add an amplifier stage to the output, but that then spoils the “pure” experience and may have an effect on the frequency response, so maybe what I need to do is find a high end USB sound card (24bit?) and put that next to the receiver circuit.

I have looked at the [[ZetaSDR|]] and I do want to build one, I am hoping that I can make this work on 30m, but I will probably make it flexible so can put in different crystals.

The long term plan is to match up a ~ZetaSDR with a USB soundcard and a Raspberry Pi as an SDR server at a remote location, and latter add a way to transmit from that location.

!! Day 2 with the design [img[]] //Shrunk down to a smaller breadboard//

Yup, I think I’ve done about as much simplification as I can here. The case means I can make sure grounding is connected correctly which I think was an issue previously. As mentioned above I have put in a bandpass filter on the input. The long cord running to my PC is probably a noise source as well. I have also exchanged the 9v battery for a 7805 regulator and running from a 12v 1.3Ah SLAB, the regulator probably doesn’t need the heatsink, it’s a left over from another project.