Technical FAQs

I-port configuration refers to whether the I and Q ports are on the top or the bottom of the package/device, and it is only relevant for surface mount or chip components (not connectorized modules, which you can just flip upside down). It allows the designer another degree of freedom in circuit layout. Most surface mount and chip components are available in our standard -2 configuration, but a -1 version could be made with a minimum buy quantity. Contact support at markimicrowave.com for more information.

Marki Microwave mixers and other devices are available with any combination of female and male connectors (all female connectors are the default catalog configuration unless otherwise specified on the data sheet).  Dash number designators for catalog models are shown below.

SMA connectors are compatible with SMA, WSMA, 3.5 mm, and 2.92 mm. Marki Microwave SMA connectors are useable to 40 GHz. For catalog products, SMA connectors are specified only for ports that function below 26.5 GHz. SMA connectors are durable, suitable for laboratory use.

2.92 mm connectors are compatible with SMA, WSMA, and 3.5 mm connectors. Mode free to 40 GHz, these connectors are useable to 50 GHz. 2.92 mm connectors need to be carefully aligned to avoid damaging the female receptacle. This makes them less suitable for laboratory use.

2.40 mm connectors are compatible with 1.85 mm V-Connectors. Mode free to 50 GHz, these connectors are useable to 65 GHz. The extended mating contact of the 2.40 mm connector is much less susceptible to alignment damage than the 2.92 mm connector, making the 2.40 mm connector more suitable for laboratory use above 26.5 GHz.

The insertion loss is the total insertion loss, not the excess insertion loss. It includes the 3 dB splitting loss. So when used as a combiner the loss will only be 3 dB, and when used as a splitter it will be 3 dB for the splitting plus 3 dB of excess loss.

Some very insightful engineers suggested many decades ago that applying a square-wave LO drive instead of a sine wave LO drive would improve the intermodulation performance of the mixer. Owing to remarkable innovations made in the design of ultra-wideband traveling wave amplifiers, it is now possible to drive the mixer LO with near-square wave signals. The bottom line is that the LO drive acts like an “ideal” switching function that opens and shorts the mixer diodes within a few picoseconds. By limiting the “intermittent” period when the mixer diodes are partially open and partially short, intermodulation distortion can be drastically reduced. For optimal results, Marki Microwave engineers have determined that square-wave LO drivers should be used with any triple balanced mixer family—especially with T3 mixers. When used with a T3 mixer, square-wave LO driving can yield an input third order intercept of +40 dBm with an LO drive around +23 to +25 dBm! 

For more information consult the T3 Mixers Video, T3 Mixers Presentation (video), T3 Mixers presentation (slides), or the T3 Mixers Primer.

The T3 mixer line is the premier mixer topology for the suppression of intermodulation products on the market today. T3 literally stands for “Two-Tone Terminator” and is especially effective when combined with a square-wave driving LO amplifier. Such amplifier-mixer combinations can be purchased under the T3A mixer line.

The answer is no we don't for the frequency and/or power they are asking for. These are residual phase noise measurements, so we had a lot of trouble finding a condition where we were able to get clean data that wasn't pushed into the noise floor of the measurement. The reason we used 6V on these measurements was because we had to use batteries to power the DC power supply to clean up the measurement enough, and so we ultimately had to chain 4 1.5V batteries together.

One of the goals that we have in the near term future is to populate our data with more information about phase noise performance vs. input power. Subsequently, we would like to be able to have some experimentation about the phase noise vs. frequency, but this is much harder to do comprehensively because of limitations of the phase noise system we have currently, and the amount of frequency-specific filtering and isolation that is needed to make clean reliable measurements at each specific frequency point.

Based on what we have seen, the phase noise performance is not linked very strongly with frequency. I would not expect the phase noise to be much better or worse at 2.5 GHz than it is at 4 GHz or 8 GHz. Where I expect it will begin changing is if you are moving to a high enough frequency that you are getting higher levels of compression at some similar input power. To first order, Enrico Rubiola shows experimentally that phase noise performance is somewhat independent of frequency on page 12 of Phase Noise in RF and Microwave Amplifiers.

The question about variation with input power is a tough one, primarily because of the difficulty in measuring reliable clean residual phase noise data for very clean low phase noise amplifiers. The question of the relationship between HBT amplifier compression and phase noise performance is one we really want to answer. Our understanding based on the literature and data we have seen is that in HEMT and FET amplifiers, there is a significant degradation in phase noise performance as the amplifier enters high compression (likely due to AM-PM conversion). We would expect a similar relationship in HBTs, but we didn't see that trend clearly when we were running our own experiments. We also found some evidence in the literature from the second paper attached that the trend isn't quite as strong in HBT amplifiers as they are in FET amplifiers. (See the final page of Jason Breitbarth's paper Additive Phase Noise in Linear and High-Efficiency X-Band Power Amplifiers).

The ADM-0126 will begin to saturate around +5 dBm input power, and will saturate completely with +10 to +15 dBm input power. This is where you should see the best T3 performance.

The saturated output power is not affected by the negative bias, so whatever value is on the curve in the datasheet should be representative.

The noise figure, however, is high at these low frequencies, around 8 dB or so.

We currently recommend using ADM-0026-5929SM or ADM-0012-5931SM.

A detailed answer to this question can be very involved, see the Mixer Basics Primer or the Mixer Basics Video for a detailed explanation. The short, intuitive explanation resides in the fact that the intermixing of tones takes place when the diodes are conducting, or at least partially conducting. Thus, there must be enough voltage drive to turn the diodes on to allow for the nonlinear intermixing of multiple tones. When the diode barrier is increased, the mixer will require a higher applied voltage to allow for nonlinear intermixing—hence intermodulation distortion products will be better suppressed.

For frequencies that low, the balun options would be the BALH-0003SMG or BAL-0003SMG. These are also available as connectorized units, although I don’t think that would be compatible with your packaging requirements. These baluns will provide the best possible suppression of the fundamental, third, and other odd harmonics. You will need to add an amplifier to counteract the loss in the baluns and diode quad.

Another option is to use a flux coupled balun from another company, if your primary concern is with loss and not harmonic suppression.

Generally, double balanced mixers have the best isolation performance due to their highly symmetrical 4 diode configuration. Marki Microwave MM1 mixers are best suited for high isolation applications, due to their computer-optimized passive structures.