Category Archives: Mixers
Everywhere you look in the microwave industry press or at the international microwave symposium you see one topic mentioned over and over again: Gallium Nitride (GaN). There is presently a gold rush in the industry to produce new varieties of products in GaN. The wide bandgap and high electron mobility of GaN mean that it is capable of a much higher power density than Gallium Arsenide (GaAs). The introduction and availability of a 0.15 micron commercial GaN processes means that fabless integrated circuit companies (such as Marki) can produce GaN microwave products at frequencies comparable to GaAs. This has led to a profusion of products including power amplifiers, low noise amplifiers, driver amplifiers, and many other amplifiers using GaN that set new records for power at high frequency.
One thing that is fueling the drive to GaN is the increasingly stringent linearity requirements placed on modern RF/Microwave systems. Since the noise floor of a system can only be reduced so much, the only way left to increase the dynamic range is to increase the power compression level and decrease the intermodulation products to create a wider spur free dynamic range. At the heart of most RF receivers is a mixer that limits the linearity and therefore the dynamic range of the receiver. So the next step looks obvious; GaN is increasing the dynamic range of many components in the system, so it should increase the dynamic range of the mixer as well!
There are five main reasons that this logic doesn’t work:
We occasionally receive requests for a mixer that will operate above our highest frequency mixer, the MM1-2567LS. The truth is that, contrary to the datasheet, the MM1-2567 actually operates above 67 GHz. It was designed to operate up to a frequency of 80 GHz on the RF and LO sides. The issue is that our test equipment can only measure conversion losses up to 67 GHz directly. To solve this problem, I devised an experiment to prove whether the mixer frequency actually extended to it’s simulation limit or not. Here is the experimental setup:
T3 mixers are the highest dynamic range mixer available. They are also handbuilt parts, subject to unit to unit and lot to lot variability. In this blog post we attempt to quantify that variability. Our sample is 10 T3-08LQP mixers from 5 different date codes. All the date codes are separated by at least a month, totaling nearly two years. Therefore, the variation you see in the plots below accurately represent the variation a designer could expect across two years in the life of their product. Of course there are always outliers, but the following represents typical performance variation.
Marki is bringing advanced mixer designs to a broader market with four new models of GaAs Schottky diode double balanced mixers covering S and K band applications. These designs combine the legendary mixer design expertise of Marki Microwave with the repeatability and economies of scale intrinsic in the MMIC production method.
The oldest question in mixer tech support is probably “what happens when I drive the mixer with X dBm LO?”, where X is some number lower than what we recommend. In general, and particularly in the past, we have avoided this question. Our recommendation was and is to never underdrive a mixer. The reason for this is that a mixer with insufficient LO drive does not act as a switching device, but as a square law device. If the LO does not turn on the diodes then the physics of mixer operation change completely, and all of our carefully laid design work is thrown out the window.
Indeed, when you underdrive a mixer the conversion loss is not the only thing that changes. All of the specs change, and in unpredictable ways. The LO side in particular responds weirdly, because a lot of deficiencies and inefficiencies on the LO side are exposed when it is underdriven and concealed under normal operation. In this post we will show all the bad things that can happen when you don’t supply an adequate LO drive to the mixer, and then leave it to you as the user to decide what LO drive to design with.
We are often asked what the ‘preferred’ method is for attaching Microlithic die to the substrate is. We use, and prefer, silver epoxy die attachment, because epoxy die attach
- is a low temperature process (epoxy can dry at room temperature, vs. very high temperatures for solder or eutectic)
- is easy to do, with a low learning curve and minimal waste of parts
- does not require a solderable die surface (not all chip based Microlithics have a nickel solder barrier layer)
- has a low probability for voids underneath the chip
- is a reversible process, where the part can be removed without damage if performed carefully.
For these reasons, we use silver epoxy (specifically Epotek H20E), place a minimal amount beneath the die, and place the chip into position so that a thin epoxy fillet is created around the chip. The part is then cured at temperature according to the epoxy instructions.
In our last post we showed the physical basis for how mixers are used as phase detectors, concluding by showing that IQ mixers make ideal phase detectors due to their ability to unambiguously identify the relative phase between two signals at any power level. In this post we examine the opposite: how to use mixers as phase modulators. It seems like you should be able to use them in exactly the opposite way, which is to apply a DC voltage to get a linear phase shift. Unfortunately, it’s not that simple.
Note: as with mixers as phase detectors, we as the manufacturers are not the best experts, but our users are. In this case I would recommend Kratos General Microwave, whose application notes I used in preparation of this blog post.
Why phase modulators?
Before examining how to get a phase modulator, let’s look at why you might need one. The main applications are communications and electronic warfare.
Communications: Phase modulation (mathematically identical to frequency modulation) has been used since very early in radio communications, due to FM communications having constant amplitude, better spectral/power efficiency, and convenience. The most common way of understanding phase modulation is with binary phase shift keying (BPSK), or quadrature phase shift keying (QPSK) if both orthogonal components are used. All modern communication systems use these techniques, so they have been written about very extensively, and we will assume that you are familiar with them.
Electronic Warfare: Here it gets interesting. If you have a phase modulator in a jammer, you can trick an enemy radar system into thinking that your plane/boat/tank is not where it actually is. You do this by listening to their radar pulses and responding with frequency shifted radar pulses, making it appear that you are moving at a different speed. This is the classic decoy technique. Modern jammer systems employ much more advanced, exotic, and classified schemes than this that I hope I never have the classification level to learn about. The principles, however, are the same.
Double Balanced Mixers as Phase Modulators
Let us start by running a phase detector in reverse. Instead of a DC output, lets input a DC signal to the IF port and a CW signal to the LO port and see what comes out of the RF port. If the device is reciprocal, then a small DC voltage/current should induce a small phase change, and a larger voltage should create a larger phase change. Here is what happens:
So this is nothing like what we expected. Why? I don’t really know. Somewhere the hand waving ‘superposition’ argument I gave in the last blog post breaks down, and something is not reciprocal. The above behavior makes sense from what we know about double balanced mixers. Namely that with no voltage applied, the LO-RF isolation prevents any signal from passing through. As the DC voltage is applied, it breaks down the symmetry of the diode quad, reducing the isolation and allowing more of a signal to pass through, although without a phase change. Now here is what happens when we apply a negative voltage:
So at least with a negative voltage you can get a 180° phase flip. So a double balanced mixer does give you phase modulation, but only between two different options. This makes it suitable as a BPSK modulator (emphasis on the word binary), but not for much else.
IQ Mixers as Phase Modulators
IQ mixers worked great as phase detectors, will they work great as phase modulators? Only one way to find out:
This picture is a little confusing, but the idea is that it works out mostly okay. As with the double balanced mixer, you see no signal pass through with 0 voltage applied. If you manipulate the DC voltage applied to the I and Q port, then you will see the phase rotate around the complete circle as expected. When you only have one voltage applied you will be almost at 0°, 90°, 180°, or 270°. There are some phase errors, I would imagine due to the non-ideality of the components, but this can probably be accounted for. There is also a non-uniformity to the amplitude due to the isolation being degraded differently in the different mixers.
So what is going on here, how can you achieve an arbitrary phase with an IQ mixer when you can only get two phases with the double balanced mixers that make them up? This is the following trigonometric identity:
a cos(x) + b sin(x) = R cos(x-theta)
where R2=a2 + b2 and tan(theta) = b/a. This means that by simply changing the values of the input signals (in this case by modulating the isolation of the mixers) you can achieve any phase within the range of the tangent function (-90° to 90°), and then by flipping the value to negative you can achieve any value in the phase circle.
Is this a good idea?
Just because you can do something doesn’t mean you should, you can drive a car with your feet if you want to but that doesn’t make it a good idea.
The short answer is yes and no. Using an IQ mixer is an easy way to achieve an arbitrary phase modulator, definitely useful if you are in the lab. A few problems though:
- The isolation change with applied DC voltage is non-linear, and the whole structure has to be carefully characterized to achieve repeatable results.
- All non-idealities are frequency dependent, so this characterization has to take place at each point in the system. Further, if you are modulating a broadband or multitone signal you won’t be able to correct for the system errors at all frequencies.
- The insertion loss when used in this way will look very little like the insertion loss shown in the datasheet for the IQ mixer. For example, here is the MLIQ-0416 datasheet conversion loss:
Nice and flat across the band around 8 dB. Now here is the same mixer as a phase modulator:
This starts out strong, but then falls off at higher frequencies. This is because the second plot is of an insertion loss instead of a conversion loss. The difference is subtle but important. In the first case the LO is only used to turn on the diodes, so the losses that it takes passing through the LO quad hybrid don’t matter very much. In the second case, the LO is the signal, so any loss that it takes shows up as an insertion loss.
Those are the main problems, and if these can be overcome, then you may be in business.
Single Sideband Mixer as a Phase/Frequency Modulator
There is another way to use an IQ mixer as a phase or frequency modulator, and this involves creating quadrature CW signals into the IF port of the mixer, effectively using the IF port as the LO, and varying the frequency of this LO signal to change the frequency offset. We will examine this, and other use cases, in our next blog post on using IQ mixers as single sideband upconverters and image reject downconverters.
Some of the most common questions we receive here are about using mixers as phase detectors. We previously discussed this topic in the post, “DC Offset and Mixers as Microwave Phase Detectors”. In this post we will go into much further depth about the physical mechanisms by which mixers act as phase detectors, and what is important for engineers trying to accomplish this in the lab. First a warning though: we’re just showing experimental results here. The real experts in phase detectors, phase noise, and all things related to phase are the people that do this every day at Holzworth Instrumentation.
Double Balanced Mixers as Phase Detectors
Much has been written about how double balanced mixers work as phase detectors (for example, see this article from Watkins Johnson about the subject). As with most circuit topics the descriptions in the literature are based in math rather than physical principles, so we’ll now consider the physical mechanisms in play when a double balanced mixer is used as a phase detector. Let’s look at what happens when we apply two in phase (frequency matched) voltage signals to an ideal double balanced ring mixer1:
The phases correspond to the phase of each signal as it appears at the diodes. Only two are show, but take my word that superposition works here and every other in between state produces the same effect. For current to flow, two conditions must be met. First there must be a voltage differential across the diode. Second is that the diode has to be pointed in the correct direction. The red arrows indicate where those conditions are met and current will flow. As you can see, when the signals are in-phase current will flow into the IF balun, creating a positive DC voltage at the IF port. Now out of phase signals:
The situation is similar for out of phase signals, except that current is always pulled out of the IF balun, thus creating a negative DC voltage at the IF port. For quadrature signals, there is equal current flowing both into and out of the IF balun. This means that no net DC current is created, no net voltage is apparent. The IF port is essentially always a DC virtual ground. This is the physical basis for why a double balanced mixer will show no DC voltage for two signals in quadrature2.
With these principles understood, let’s go in the lab and see what happens when we actually apply these signals to double balanced mixers. First we create two signals and use an oscilloscope to verify that they are in quadrature:
At 10 GHz the period is 100 ps, so 25 ps out of phase is in quadrature. Now we apply these voltages to the input of a ML1-0220LS mixer, and what do we find? Nonzero voltage! In fact, here is what the DC output voltage (taken with a terminated bias tee, this is very important) looks like a function of phase between the two input signals:
Now we would expect this to be a peak at 0°, and the minimum to be at 180°. What is going on? This is a phenomenon that is documented by Stephan Kurtz in the previously referenced WJ app note. In modern double balanced mixers the RF and LO baluns are not identical. In fact, they are not even close. One side is built as a magic tee, where the IF is removed, and the other side has a return to ground on it. Even though the LO and RF baluns traditionally cover identical frequency bands, there is no reason that they need to. They can be completely different! This means that they most likely have a different electrical transmission length and phase delay, which is why the peak of the voltage curve is not quite at 0°. Another effect highlighted in this app note is that there is a voltage offset that shifts the entire curve up (or down). As we detailed in our first post, the excellent balance and isolation of the ML1-0220 minimizes this DC offset and makes it not noticeable for this plot.
1 Note that the necessary DC current return to ground path is necessary but not illustrated for clarity.
2 It is easy to imagine how to extend these principles to the situation where the signals are not at their peak or zero values, and similarly to phases that are not either perfectly in phase or out of phase. While superposition does not strictly work in a nonlinear system such as this, the results one would expect from superposition are maintained qualitatively.
IQ Mixers as Phase Detectors
Now we can calculate the phase of the signal. Excellent. However, there are two ambiguities that we need to clear up. Since this is a sine wave instead of a sawtooth wave, there is some ambiguity about the phase. The same output voltage could be two different phases, except for the max and min. This is fine if you are doing phase noise testing, where you put the two signals in quadrature and just look at any voltage that comes out. For actually detecting the phase between two signals though, it isn’t enough information. The second ambiguity is that we need to know the max and min voltage levels, as well as the DC offset, to determine the phase. Since the DC offset in Microlithic mixers is small we can ignore this, but we still have a problem if the incoming signals change power at all.
How do we resolve this? One way is to use two mixers as phase detectors and deliberately introduce a phase shift between the two inputs (RF and LO). 180° is no good, because the phase ambiguity remains, so a balun is out. A length of line changes phase with frequency, so that is out too. The other broadband phase shifting options we have are a Schiffman phase shifter or a quadrature hybrid. The quad hybrid is much more common and easy to build3, so what would a structure with a quadrature hybrid introduced on one side look like?
That’s right: the structure is exactly an IQ mixer. Since I and Q are in quadrature, it is easy to calculate the phase between the two signals as
after making a small correction to scale the I and Q values by their peak output level and DC offset4. Let’s look at the same plot of voltage vs. phase for the IQ mixer, along with the calculated phase:
As expected, the calculated phase is almost linear with input phase after the correction factors. This is a significant improvement over the double balanced mixer, since we don’t need to know the input power levels and there is never any phase ambiguity. But how close is the IQ mixer to invariant with input power? When the two signals are at 0/90/180/270 degrees to each other, there is obviously very little variation in calculated phase with power since one of the voltages doesn’t change. If we pick a phase in the middle (135°)5, this is what it looks like:
As you can see the power levels the agreement to 135° is excellent. As we increase to higher power levels, one of the mixers compresses sooner than the other mixer, and the phase is thrown off. Up to 0 dBm, however, the agreement with the real phase is excellent. This does not address what happens when one of the signals is significantly higher than the other one, nor with double balanced mixers when you are just trying to detect phase changes, where high powers are desirable to increase sensitivity.
Now that we have examined the physical mechanisms of how mixers work as phase detectors, we can do the reverse and see how they work as phase modulators. This is the subject we will tackle in our next post, “All About Mixers as Phase Modulators”.
3Quad hybrids are easier than Schiffman phase shifters, but still ridiculously difficult to build broadband. You don’t have to trust me, you can try yourself, and then buy ours when it takes you 6 months.
4You also have to convert to a -180 to +180 phase range, or 0 to 360, or whatever. Arctangent only gives you values from -90 to +90, so you have to use the sign of the signals to figure out where exactly you are.
5How do we know the phase is 45 degrees? Because we put the signals in quadrature (which is the same at any power levels), and then moved them 12.5 ps on the oscilloscope, equivalent to 45° at 10 GHz.
Spectral regrowth is a big deal for you. In order for the wireless revolution to continue apace, enabling you to watch funny cat videos faster in more crowded environments, spectral regrowth must be conquered wherever it occurs. Spectral regrowth is what occurs when a broadband or spread spectrum signal intermodulates with itself, creating deterministic products that look like noise, limiting the signal to noise ratio of the signal. According to Shannon’s theorem this limits the information capacity of the signal, and thus your video takes longer to load (for some reason this always happens at the worst time).
Spectral regrowth comes from a handful of sources. It can come from mixers, but in installed communication systems it tends to come from the power amplifiers at the transmitter and the connections to the antenna itself (called ‘passive intermodulation’ or PIM). It is made much worse by using higher power and by denser concentrations of signals. Both of these factors are increasingly common as data capacity is increased. This is why highly linear amplifiers and PIM are both big buzzwords in the mobile communications world right now.
What is not always talked about is that ‘spectral regrowth’ in the mobile communication world is the same as ‘two-tone intermodulation distortion’ or ‘IP3’ in the microwave world. Two tone modulation distortion is what causes spectral regrowth, just summed over all of the frequencies involved in the signal. This can be seen by moving from two tone testing to three tone testing in a standard double balanced mixer, the Marki M1-0212SA:
As you can see, it is a complete mess! This is considered very good for a double balanced mixer, and it is better than you would get from any GaAs mixer on the market. This is with 0 dBm output signals and a 25 dBm LO drive (square wave, this will matter in a minute). Note that we use 0 dBm output power as the reference instead of the input power. As we have mentioned before, this is what you really care about as a designer (how much range do I have at the final stage in my receiver) and accounts for the variation in insertion loss between mixers. You can cheat by adding loss to a mixer to improve the input IP3, but you can’t cheat on the output IP3.
This output spectrum is obviously unsuitable for operating with, nonetheless testing a high performance system. The testing system must have at least 20 dB more dynamic range than the system itself, and usually much more, so getting rid of these obnoxious intermodulation products is critical. This was the state of the art for many years, then Watkins Johnson came out with their termination insensitive mixer, the M8TH (still on sale from Ma/com and still considered the standard by some). Here is the M8TH output spectra:
Much better! The intermodulation products have been suppressed significantly, although they are still significantly limiting the dynamic range to 45 dB. Recently there has been much talk of the FET mixer. At Marki we have not been that impressed with the FET mixer, because of the narrow bandwidth over which it works. Combined with the intrinsic poor isolation available in the FET circuit, this limits the performance below what we like to see. Nonetheless people get excited about it, so here is the narrow bandwidth, poor isolation PE4140 FET quad mixer we built two and three tone results:
Good improvement above the termination insensitive mixer, now the dynamic range is up to 50 dB. At least I guess it is cheap and good for communication applications, but with the narrow bandwidth and isolation problems it isn’t well suited as a test and measurement mixer. Now let’s look at the T3. As a true commutating mixer it is in its element with a 25 dBm square drive. Here are the results:
And this is also the mixer with the best spurious suppression, and isolations and conversion loss as good as any mixer available. All of this across a 2000:1 bandwidth ratio. It isn’t fair to say that the T3 is the Ferrari of mixers. It’s more like the T3 is a fighter jet racing a car. When it comes to choosing a mixer for your test system, there really is no choice but the T3.
Every day we work on high linearity mixers: high IP3, high P1dB, and high spurious suppression. Every once in a while we get a request for a high IP2 mixer. This is much more rare than complaints about IP3 or spurs. Lets see why.
To start understanding IP2 in mixers, lets look at intermodulation products in amplifiers. Start by imagining a single tone into a single non-ideal amplifiers, with a nonlinearity.
A single input tone is amplified to a larger tone, as desired. Due to the nonlinearity in the amplifier, higher order tones are created. These are only created at integer multiples of the input tone, since the system is time invariant. These decrease in power as the frequency increases, and they are generally referred to as second, third, Nth order distortion products.
A new problem arises if we put two tones into the same amplifier:
This diagram only shows the second harmonics, the second order intermodulation, and the third order intermodulation in the relevant bands. As you can see, the second harmonics (at 2f1 and 2f2) are easily filtered out as they are at a very high frequency, unless you are in a very broadband multioctave system. The second order intermodulation (at f1 + f2) is in between the two harmonic distortion products, so it is also easily filtered except in broadband systems. While these are drawn as the same power level, they are not necessarily the same power in practice. The third order harmonics (at 2f1-f2, f1+f2-f1, f1+f2-f2, and 2f2-f1) are all famously in band. Two of the tones are directly on top of the desired received tones, and therefore they cannot be filtered even with an infinitely narrow and steep filter.
Now to mixers. The addition of the time varying LO increases the complexity dramatically. All of the previous effects will be present, and also will all be present at the LO crossed frequencies. For the moment lets ignore everything but the isolations, fundamentals, and second order distortions (both single and multitone).
In this case, there are two types of second order products: direct and converted. The direct products are at the same frequency as in an amplifier, and the converted second order product is between the 2IF x 1LO products. Again these are drawn as the same power level, but they may or may not be identical powers.
When will these products matter? The converted products will appear in the passband in a broadband system (with a low IF) where they cannot be filtered out. Conversely the direct second order products will matter in a different kind of broadband system, with a high IF. Specifically when fIF = 1/3 fLO, the converted signal and the second harmonic of the IF, and the second order distortion product, will all be at 2/3 fLO. These can be a problem, but usually no worse and closely affiliated with the 2IF x 0 LO spur.
So in both cases we see that while the second order distortion exists, it is always close to a high power spur that also must be dealt with in the frequency plan.
Now lets consider a downconversion:
Once again the converted second order shows up, this time in between the two 2LO – 2RF spurs that usually wreak havoc on downconversion systems. Once again the same frequency plan is needed to eliminate it. The direct second order term, however, is at f2-f1, which becomes a significant problem when the IF frequency is similar in magnitude to the separation between the two tones. In this case the direct second order tone would lie directly over one of the tones.
The converted two tone second order intermodulation product will be an issue in the same circumstance as the 2x-2 spur is a problem, namely when you have a low IF. If the IF is at DC (direct downconversion) then the second order intermodulation will cause significant distortion at DC. This is why the most common reference to IP2 in the literature is for the mitigation of it in direct downconversion receivers.
One thing that is not IP2, but is sometimes referred to as IP2, is the half-IF spur. This occurs when a signal (at Frf) is downconverted to a low frequency, near baseband, and there is a jamming signal at a frequency (Fj) roughly halfway between the RF and LO frequencies. The downconverted jamming signal can be filtered out by the IF filter, along with all other unwanted signals. However, the jammer signal creates a high power 2 LO x 2 RF spur, however, that will show up at or near the desired signal, and there is no physical way to filter it out.
Fortunately, in either a double or triple balanced mixer structure the 2×2 spur will be well canceled by both the LO and RF baluns, resulting in excellent suppression when well balanced baluns are used. For example, the ML1-0218ISM offers a downconversion 2×2 suppression of 58 dBc with an input of -10 dBm. The T3 circuit can offer even better suppression, since the proprietary T3 circuit will both prevent and suppress these spurs. Therefore the T3-18 offers a superior 64 dBc suppression of the 2×2 spur with the same -10 dBm input.
However, this is not a two tone IP2 problem. It is simply a second order distortion product. So I shouldn’t take these authors too hard to task. As Joel Dunsmore cautions in his book, Introduction to Microwave Measurements:
There is sometimes confusion in the use of the term second-order intercept; while it is most commonly used to refer to the second harmonic content, in some cases, it has also been used to refer to the two-tone second-order intercept, which is a distortion product that occurs at the sum of the two tones. Most properly, one should always use the term two-tone SOI if one is to distinguish from the more common harmonic SOI.
And that is the final point of this post; when you are talking about IP2, you always need to be specific about what you mean.