RF Mixer Levels

[dohzer]

I'm building my first superheterodyne receiver, and I've done a bit of reading, but one thing I've never really seen mentioned is what range of signal levels I should be feeding into a mixer?

Does it matter? Should I be trying to target equal RF and LO levels?
Part of me thinks that it probably doesn't matter as long as the frequencies are correct, but is there a rule of thumb for such designs?

[German_EE]

First thing head over to the Minicircuits website and have a look around. http://194.75.38.69/homepage/homepage.html Yeah, I know that it looks weird using an IP Address like that but their I.T. Department has issues.

So, click on 'Products' then 'Frequency Mixers' and you'll see things like "Level 3-4" and "Level 7". This indicates the level of the Local Oscillator signal in dBm that the device should be fed with but ONLY at the local oscillator port.

Now, here's the trick. A double balanced mixer contains a matched set of four diodes. Your Local Oscillator signal should be strong enough to turn on alternating pairs of those diodes in normal operation but your RF input signal should NOT be strong enough to turn on the diodes. This normally means that your RF input should be at least 6dB below the Local Oscillator signal.

Confused by dBs and dBm? Send me a PM with your email address and I'll let you have a copy of a chart I created that converts between dBm, RMS volts and Peak-to-Peak volts.

[dohzer]

Wow thanks for the concise answer!

Now, here's the trick. A double balanced mixer contains a matched set of four diodes. Your Local Oscillator signal should be strong enough to turn on alternating pairs of those diodes in normal operation but your RF input signal should NOT be strong enough to turn on the diodes. This normally means that your RF input should be at least 6dB below the Local Oscillator signal.

I'm actually using a minicircuits mixer; an ADE-1+. The LO level should have been fairly obvious from the datasheet... my fault. But that's a really good guide to the RF side of things. I'm finding it a little difficult to measure signal levels without RF equipment (I was actually surprised when my (hacked) 100MHz Rigol scope told me my signal was a 159MHz), but I'll build a peak power detector when my diodes arrive in the mail.

Confused by dBs and dBm? Send me a PM with your email address and I'll let you have a copy of a chart I created that converts between dBm, RMS volts and Peak-to-Peak volts.

I get the whole conversion side of things, although it does take a bit of mental gymnastics when starting out. And ensuring that I stick to the right impedances.

[vk3yedotcom]

I'm building my first superheterodyne receiver, and I've done a bit of reading, but one thing I've never really seen mentioned is what range of signal levels I should be feeding into a mixer?

Depends on the mixer.  A diode balanced mixer will have a much higher requirement (maybe 5 - 10 mW) than something like an NE602 or FET which needs very very little. 

However the diode balanced mixer will have better stronger signal handling performance.

[rfbroadband]

you drive the LO levels usually with more than 0dBm. The key parameter that often determines required LO levels is non-linearity (IP3), not sure if you have considered this. The levels at the RF input depend on your required dynamic range. Usually you proceed  the RF port with a variable gain LNA to be able to get closer to the noise floor of your system. The max. gain of the LNA will introduce a limit on the highest level that can be used at the RF port (again non linearity).

[T3sl4co1l]

Drive levels depend on what the mixer actually is, and how much signal it needs.

...Duh?

Well...

If it's a DBM type, then it needs enough LO power to forward-bias the diodes.  To behave as a toggle switch, this must be at least a few times the maximum RF input (I believe +10dB is typically suggested), and needs to be from a low enough impedance to drive the diode capacitance at the desired frequency.

Note that the source impedance is used as current limiting; all the LO is doing, is driving current through some diodes.  So as those diodes conduct, they will reflect a lot of power.  So, expect a poor VSWR.  Which is confirmed with the datasheet.  Also expect more drive power (and perhaps a lower LO impedance) for higher operating frequencies.  Plan accordingly!

If it's a transistor type (e.g., CA3028 single balanced; MC1496 fully balanced; anything JFET or MOS, from single or fully balanced bridges to analog switches), then the goal is again to deliver enough voltage to achieve good mixing, while using a low enough source impedance to drive the capacitances to that frequency.  But, there is no dependency on RF input power (or at least a lot less).  The drive power can be much lower in these cases, which you can imagine as obtaining amplification from the transistors, or at least getting light loading from FET gates.  BJTs only need ~0.1Vpk to nearly fully switch on/off, with junction capacitance similar to diodes'.  FETs are mainly capacitive, so the resistance can be quite high, though if the capacitive reactance has to be canceled with inductance, the LO frequency range might not be wide (BW limited by narrowband tuning).

Note that "system impedance" is simply what's handy, and what they measured it at.  YMMV, for better or worse, if you wish to try other impedances.  If you can't take the time, follow the datasheet; if you can, there are probably gains to be had (power efficiency, noise floor, dynamic range).

Tim

[AF6LJ]

This brings back memories; when I worked at Loral Corp. we used the CA3028 as a mixer in our microwave transmitters it made a great mixer and went easy on the current drain.

[Wolfgang]

Hi,

if you want to use a DBM the mixer level you need is determined by the IMD requirements you have. In the heydays of Schottky mixer frontends, the professional receivers of Rohde and Schwarz used mixers up to +27dBm of LO power.

Nowadays even better IMD can be obtained by FET switching mixers (see EMRFD).