0
\$\begingroup\$

I am self teaching RF via the internet, "Introduction to Radio Frequency Design" by Wes Hayward and the 2005 ARRL handbook. My intended project is a 1 stage IF heterodyne receiver. Not following any recipe, just trying to apply the topics that I learn. The mixer as shown has significant insertion loss, which I am looking to minimize.

The build:enter image description here

Schematic:enter image description here

-1N34A Ge diodes with ~0.3V Vf.

-I change the LO frequency by sliding a ferrite core into the LO coil. Range is 5 MHz to 11 MHz.

-Toroid is a FT37-43 core with a trifilar winding. Zprimary = 4 times RF or LO system impedance at 500 Khz, my lowest frequency of interest. That came out to approx. 120 ohms and 10 turns, but I could only fit 9 turns.

-Termination occurs via a 47 ohm resistor to ground. I suspect this is the largest contributor to insertion loss.

-I want to up-convert (0.5 MHz - 1.8 MHz) to the neighborhood of 8 Mhz, then use an 8 MHz crystal to filter out my IF for amplification. The crystal would be in parallel to the 47 ohm resistor to ground and lead into an IF amplifier. Envelope detector demodulation and audio amplification would follow.

-If I put my finger on the antenna (I become a larger antenna than my random wire), my RF and IF peaks rise. I was thinking to turn my RF amplifier into a darlington pair to increase my RF, but I would like to improve what I currently have before that if possible.

Probe Tip:

LO Output

LO Output (pad with the black mixer wire probed): Tip

Mixer (Big squares were probed): MixerInput

Mixer (Small squares between diodes and resistor to gnd were probed): MixerOutput

Mixer (small squares between diodes and resistor to gnd were probed with me acting as antenna): enter image description here

Are there any design suggestions that I can implement to minimize insertion loss and get higher amplitude IF peaks on my mixer output?

Edit:

Probing RF output (pad with green and red wire to TR) reads -24.5 dBm (assuming 10k impedance, lines up with the 530mv p2p displayed too).

Probing Mixer output (small pads between diodes and resistor) reads -35 dbm. Loss of 10.5 dBm isn't so bad I guess.

-I will feed the RF output into another identical emitter-follower to increase RF drive with a smaller antenna and maintain Zout of RF.

-IF will be filtered with a single crystal that will lead into an IF amplifier.

\$\endgroup\$
6
  • 1
    \$\begingroup\$ Please edit your question. First, what is your "significant loss"? If you're not sure, guess. I know that an ideal passive doubly-balanced mixer has about 6dB of loss. I'm pretty sure that an ideal singly-balanced mixer is going to have around 10dB of loss. Second, why do you feel that loss in a known-lossy mixer topology is a bad thing? In general if you're using a diode mixer you plan on the loss and use an RF preamp at frequencies above 7 or 10MHz, and you use a buffer amplifier between the mixer and the IF filter. What are your plans for those amplifiers? \$\endgroup\$
    – TimWescott
    Jul 29 at 18:46
  • 1
    \$\begingroup\$ With this single-balance (two diode) mixer, half the signal is thrown away. The 4-diode version, with another toroid gives you more. Also, have never seen 1N34 diodes in that type glass package. That preamp may be weak when faced with monster AM broadcast signals - it likely needs much more DC current to stay linear. \$\endgroup\$
    – glen_geek
    Jul 29 at 18:49
  • \$\begingroup\$ You probably want to beg borrow or build a signal generator, even if it's just an oscillator. You probably also want to make it well-shielded, and build yourself an RF attenuator so you can inject a known signal into your antenna port. \$\endgroup\$
    – TimWescott
    Jul 29 at 18:49
  • \$\begingroup\$ On @glen_geek's comment about the diodes -- I'm not sure what motivates the choice of germanium diodes. The only thing you lose by going to silicon diodes is the need for a higher-level LO signal -- you don't cause much more loss as long as the diodes are turned on all the way. In return, you gain a bit more IMD performance, and you can use much more commonly available diodes. \$\endgroup\$
    – TimWescott
    Jul 29 at 18:52
  • \$\begingroup\$ @TimWescott Thanks for the context of expected loss in a diode mixer, that put things into perspective. I am now fine with the loss given how improvised my setup is. I was originally planning on a single crystal filter in series with my mixer output then amplifying my IF, similar to your buffer suggestion. I had the Ge diodes sitting around from a crystal set I made a while ago, no specific motivation for those. Feel free to post your comment as an answer. I will confirm these results with the signal gen built into my oscope. \$\endgroup\$ Jul 29 at 20:10

1 Answer 1

1
\$\begingroup\$

Your mixer is lossy because of how you have configured your ports:

enter image description here

With the LO feeding the tx primary, then both diodes conduct on one LO half cycle (if the LO power is sufficient), and neither on the other. I'm not sure if this classifies as a balanced mixer, but one thing is sure, the output is 6dB down on what it could be if both half cycles are used.

The more usual way of configuring this single balanced mixer is to swap your LO and RF ports:

enter image description here

then the RF signal is available as 0/180deg at S1 and S3, and these are switched to the IF load on alternate half cycles of the LO.

e.g. see: edadocs.software.keysight.com/display/genesys2010/Diode+Transformer+Single+Balanced

Keysight suggest a typical conversion loss of 7dB.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the insight and the great resource! \$\endgroup\$ Jul 29 at 23:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.