Unexpected/unwanted output on 1GHz mixer

[nealmartini]

My circuit has a two Minicircuit VCOs driving the RF and LO ports of a Minicircuits mixer. The VCO signals are very clean when tested separately, but when I hook them up to the mixer and connect the IF port to a spectrum analyzer there are all kind of unwanted components. The IF is at 1Ghz. I need suggestions/references on what I need to do to get rid of the unwanted stuff at the IF port. For example, should I be using pads to help improve VSWR? Should I be filtering inputs and outputs? Help!!!

 

[logbook]

Unless it’s secret, why don’t you spell out the signal levels, frequencies and component type numbers you are using. Also spell out what "all sorts of unwanted components" means.

By knowing the frequencies of the spurs it is possible to work out what they are and where they are coming from. When you say the VCOs are very clean, what does that mean? Are the harmonics down at less than -70dBc?

Over or under driving the mixer is going make the situation worse. Using a pad will improve the matching, and reverse isolation, but reduce the signal. If the signal was the right size for that mixer, padding it down may make the situation worse. The "trick" is to use the mixer as the manufacturer intended, ie with the correct signal levels.

 

[nealmartini]

My application requires mixing at 1.5GHz. I am currently just experimenting with the MiniCircuits parts to become familiar with RF mixing.

I am using a MiniCircuits VCO ROS-1500 (8.7dBm)for the LO drive. I am using a ROS-2160W (5dBm) as a wideband source.The mixer I am using is a SYM-25DLHW. The VCO levels match the mixer pretty well.
I have the output of the mixer direct feeding into an HP spectrum analyzer. Unfortunately, it only goes up to 1.6GHz.
When I say the VCOs are clean, I mean there are virtually no spurs in the 0-1GHz range that I can see on the spectrum analyzer. The miniCircuits spec sheets show the harmonics down approx >=10dBm from the fundamental.
The mixer RF and LO VSWR are about 2:1 and 1.2:1 respectively. I don't have enough experience in RF to know if directly connected VCOs with these VSWRs is acceptable. As you point out, pads and/or filters would reduce the levels which in the case of the LO, would be a problem.
I also am using a "homemade PCB". I have been very careful to have a good groundplane and trace widths to yield a 50 ohm impedance, but it is hand soldered. Once again, I don't have enough experience with RF to know if this OK.

Finally, your question about the spur frequencies. The spectrum from 0-1Ghz is where I am looking. There are maybe 4-8 spurs that move around as I vary the VCO tuning voltage. There levels are 15-18dBm down from carrier.

Sorry this was so long, but I hope this paints a clearer picture! Any suggestions?

 

[krytar]

Hi,

Lower your RF power will help to lower the spur level. Put a 10dB pad between RF source & your mixer.

Usually, the mixer need very little power from RF to activate the mixing process. LO is the main drive to turn the mixer on. Some mixers can turn on with RF at -20dBm, but most of the time, -15dBm to -10 dBm is the required power.

I hope this will help!

Krytar

 

[davicente]

Hi,
apart from redducing input power levels, try filtering after mixing as spurious from both LO and RF signal could be producing undesirable mixing frecuencies

 

[Higgler]

I've seen alot of "applications" info.
Check Minicircuits for "mixer application notes". Google it too and add the word spur or spurious signals in the search.
Power level settings, improving the output (and input)match of the mixer helps, then you're only left with filtering. I think typically that you can't expect any real purity from a mixer output since a mixer is an on/off switch which automatically produces alot of harmonics of the input signal.
Good luck,
kch

 

[biff44]

Download the spur calculator program from here:

http://www.oml-mmw.com/calc.htm

Set Mx=9 and Nx=9, and see how many spurs are mathematically predicted.

In the real world, the level of the RF input port will affect the spur level of SOME of the spurs.

 

[zeitghost]

There used to be an ap note "Care & Feeding of Double Balanced Mixers" on the Miniccts site.

All about putting a diplexer on the IF output.

 

[MKimagin]

it might be several reason why the signal get unwonted components, let see:
If the LO VCO and RF VCO using the same VCC and Vtune supply line the signal will lick thru that line, from one VCO to other good filtering of that line is critical.
The isolation in mixer between LO to RF and LO to IF is not that big, so if the LO signal is like 10dBm and the isolation is 20dB on RF port then the RF VCO getting substential signal that can not be ignored. Mutching also is a problem. To make some improvements of RF port u can use monolitic amplifer to bust the signal and use atternuator in series with that Amp , that give u 0 gain but improve isolataion and matching significantly (all spectrum analyzers use similar approtch on input port- very effective).