I have been a practicing RF engineer for many decades. One thing that has always interested me is the criteria other engineers use to buy and use test equipment for their labs. It is not uncommon to see an expensive HP/Agilent/Keysight or Rohde and Schwarz signal generator in a lab, set to the same common frequency and amplitude for years. The old saying: “Nobody ever got fired for buying IBM” seems to apply here, especially at capital equipment budget time.
 
I have also noticed that more important than fancy modulation and all the other bells and whistles, was the need for more than one signal source used in a test setup. Generally, something like two local oscillators, or a local oscillator and sampling clock, or maybe two RF test tones was always a common requirement. These sources didn’t need to be anything particularly special; they just needed to be available. That is what prompted us, several years ago, to start designing and manufacturing our own line of modular signal sources that could replace expensive laboratory test instruments.
 
Our market model has always been to provide a low-cost signal source module with good signal purity and stability that was easy to program and simple to integrate into a test rack or equipment chassis. Non-volatile data memory has been a key function in every synthesizer we made. We always had models with USB communication, but we shied away from ever requiring USB power.

The new 15GHz Valon 5015 synthesizer  is a culmination of our current understanding of our customers’ needs and budget. Frequency range was one area for improvement. The Valon 5009 went to 6GHz, which continues to satisfy many of our customers’ applications. However, we became aware that a few of our customers were using our synthesizers with frequency doublers and triplers to get into the higher bands. So, the Valon 5015 was developed with the specific goal of covering all of X-band and most of Ku-band without sacrificing phase noise. In fact, the 5015 phase noise is outstanding. Take a look at the plot below. That’s -100dBc/Hz at 10kHz offset at 15GHz, which is very good for a wide-range signal source for under $2K.

Providing higher frequency output is only useful if you can also provide higher power. That is why we wanted to get at least +10dBm across the entire microwave bands from 1GHz to 15GHz. The 5015 is guaranteed to provide greater than +13dBm from 100MHz to 15GHz and typically can supply +15dBm at most frequencies. That is a lot of power particularly in the microwave range. That means you can drive many devices like multipliers directly without any additional power gain stages.
 
Of course, it is not much of a benefit for a source to go to 15GHz if the phase noise is mediocre. So providing the best phase noise at a very low price point was also a top priority goal. To keep the price under $2000 we had to rule out using YIG tuned oscillators, multiple offset loops, and ovenized crystal oscillators. Instead we evaluated all manufacturers of integrated synthesizer chips and various topology configurations using our signal source analyzer in order to select the best devices. By using the best available ICs with our offset sub-synthesizer topology, we were able to achieve better than -100dBc/Hz at 10kHz offset at 15GHz, which is very good.
 
Providing low spurs and avoiding boundary spurs was a secondary goal. The Valon 5015 does use a fractional-N main synthesizer so “frac” spurs are not completely unavoidable, but they can be greatly reduced. The 5015 topology uses a somewhat common method of changing reference frequencies in the 100MHz range to dodge boundary spurs. What we do that is unique is our method of generating a low phase noise sub-synthesized reference frequency and the reference frequency selection algorithm. This method minimizes boundary spurs while not sacrificing phase noise.
 
In addition to the low phase noise and high power, another innovation we added in the 5015 is calibrated absolute output power. Two new factors contributed to this development. One is the use of the best broadband, flat-gain, MMIC amplifiers. These are expensive but we feel they are worth the cost. The other innovation is our automatic synthesizer test station. This is an amazing combination of test hardware and software that allows us to automatically test output power at many frequency points and power levels and record the results in a test file for each synthesizer. The synthesizer then uses this data file to interpolate and compensate the synthesizer output for the exact power output at any particular frequency. This new scheme has the benefit of not only allowing the user to specify a desired output power in dBm accurately, but also allows for fine amplitude level setting. The overall result is a flat and calibrated output power over the entire 10MHz to 15GHz range.
 
All of our synthesizers have provided a USB control interface so the user can use either a command line based terminal emulator program or our GUI to talk to the synthesizer. The 5009 provided a second asynchronous 2-wire TTL port as another serial interface to our synthesizer. This is great for most machine-to-machine control applications, but some of our customers were using multiple 5009s in large arrays for research and factory control applications. The USB or serial interface was not the optimum way to easily control many devices.

So, the Valon 5015 adds an Ethernet access port as well. We didn’t include the typical Ethernet RJ-45 in our product, but instead used a high quality 2-mm connector. The advantage is that the housing can be much thinner. This means less wasted material when machining, but it also makes the EMI shielded cavities smaller and less prone to moding. This is common in a microwave device that covers multiple octaves. The other consideration is that including the RJ-45 in our product would limit the ability to embed our product in our customers’ application. A lot of our synthesizers end up in a rack-mounted chassis which contains a lot of other hardware. Having the ability to put the RF end of the synthesizer where it is needed in the most direct way while still allowing the Ethernet connection to be located remotely is a huge advantage in many applications.
 
Lastly, I’d like to mention our Valon 5015 switching power regulation. Many engineers think that a switching power supply has no place in a low-noise RF product. But if that was strictly the case, there would be no spectrum analyzers, no vector network analyzers, and virtually no other pieces of commercial test equipment. A synthesizer as advanced as the 5015 uses more power than can be handled with linear regulation. Plus, we didn’t want to limit our customers to using a narrow power voltage range. We have a lot of experience designing RF products for our customers that integrate an RF signal processing scheme in amongst a lot of power-hungry digital components. So we know how to make switchers pretty quiet. The switching frequency and its harmonics in the 5015 are virtually undetectable. We are pretty proud that the 5015 can use any input power from 5Vdc (4.5V guaranteed minimum at the dc input connector so you can have some cable loss from your 5V supply) to 15V maximum (and a skosh over, 16V maximum).
 
By the way, like all Valon products, all our devices are reverse voltage protected. Even with an industrial component like ours, it is important to realize that mistakes happen. No user wants to have his project put on hold because he accidently blew up a device due to the power supply connection being configured ass-backwards.
 
The Valon 5015 incorporates many years of RF experience and know-how into a new 15GHz synthesizer module. We are sure you will find that it offers precision and versatility at a low cost, and look forward to hearing about how you integrate it into your applications.

Phase noise however is one of the most difficult of all electrical phenomenon to measure.  Typically, it can be observed using direct measurement techniques with spectrum analyzers. Some advanced spectrum analyzers will even have a built-in phase noise analysis function.   However, even the best spectrum analyzers have limitations on making close-in, low-level phase noise measurements accurately. The problem with making phase noise measurements using a spectrum analyzer on precision signal sources is not knowing what the contribution of the spectrum analyzer is on the observed measured phase noise.  Or said another way, how do you know if the phase noise you measure is your measurement instrument or the DUT (device under test)?
 
Fortunately, some brilliant mathematicians (magicians) and electrical engineers (wizards) figured out a system of measurement using a technique called “cross-correlation” that systematically removes the noise contribution of the measurement equipment.  Genius!   Unfortunately, these boxes come at a hefty price.  They are often as expensive as the best spectrum analyzers in the same frequency space.
 
Here at Valon Technology we are committed to providing the best products we can, so we invested in this cross-correlation technology.  Having this test tool on the bench allows us to optimize our product design.  We discern such nuanced effects such as how a voltage regulator can affect the phase noise or what is the optimum signal level to drive a phase detector (high) to get the best performance. 
 
Our engineering bench uses the Berkeley Nucleonics 7300 series Signal Source Analyzer to help us ensure we are squeezing all the squeaks and fuzz out of our frequency synthesizers and dividers. That is why we stand by our product performance.

I’m past the age where most engineers would refer to me as old school. I’m nearly Jurassic. I still have HP spectrum analyzers on the bench. But over the years I’ve learned a few things about debugging and bringing up a new hardware design. 
 
Recently a young engineer working for one of my clients was having trouble getting her microwave PLL to lock.  She asked me what could be wrong and showed me her loop filter calculation with all the pole-zero calculations carefully plotted. Her concern was that her stability analysis was somehow wrong and that was what was causing the no-lock problem.
 
Microwave oscillators and PLLs can be a bit daunting and the textbook theory can be complex. My experience with debugging is that it is usually something basic and simple that is wrong. In most cases you can usually find the bug with nothing more than VOM.  Furthermore, almost all phase lock loops will lock even if the loop filter is way wrong.
 
I asked her if she’d checked all the IC pin voltages in the associated circuit and she said she had not but assumed that wasn’t the problem since the power regulators were operational. She came back with a still grim look and said all the voltages were fine and wondered if she should replace the chip. I asked her to check the PLL reference clock to see if it was present. She came back this time with a big smile and told me the clock was present but it was 100MHz and her PLL reference divider was programmed for a 10MHz reference clock. When she changed her code to set the reference divider correctly, the microwave oscillator snapped to the correct frequency and LED lock light came on.
 
The point is:  Stay calm, check the simple stuff first, this too will pass.

Figure 1 Lab bench power supply can cause problems if you "dial up" the voltage.

The power supply is not the problem nor is there anything wrong with the 5009. The problem is how the power supply is set up and used.  Follow these suggestions to make sure you don’t run into start up problems:

1. Current limit set too low. Make sure the current limit is set well above the required 560mA. Best to set the current limit to 1.0A.
2. Preset the output voltage to 6.0+V. Don’t set the power to exactly 6.0V. It’s best to set the power supply a little higher in order to compensate for I*R drop of the power supply cables.
3. Use the power supply on/off switch to power on the 5009. Biggest problem with erratic 5009 performance with lab supplies is the user “dialing up” the voltage slowly from 0V. This will sometimes confuse the power-on-reset circuitry resulting in an unlocked synthesizer. It’s best to just adjust the output voltage first and then power-cycle the power supply or plug the power cable into the powered on power supply.

Hope this helps.  Above all, if you have trouble or need help just contact us.

The Valon 5009 is guaranteed to go to 6 GHz. But in an emergency you can go a bit beyond, actually quite a bit beyond. The 5009 has an undocumented command that we use for testing that enables us to see how much margin is available to the tuning range of our VCOs. You can use it too. Just type the command “debugflag 1001” and the 5009 will ignore the preset frequency limits. In most cases you will be able to go beyond 6.2 GHz on the high end and approximately 1 MHz lower than 23.5 MHz on the bottom end.

Remember, you’re operating out of range so don’t call us if you can’t get to 8GHz using this special command. We just thought you might like to know it was there.

The 5009 provides two serial interface options for the user. One is the standard micro-USB port that provides a convenient connection to any PC. The other available serial port is the USER port provided by the 8-pin Hirose 2mm connector next to the USB interface. The USER port has two pins, RXD and TXD (and ground), that are intended to connect directly to the users host such as an Arduino, Raspberry Pi, or embedded microprocessor. RXD and TXD are 3.3V TTL logic signals.

Note that both serial ports can be used simultaneously; there is no conflict and neither port has priority over the other. Two separate terminals or controllers can control the same synthesizer and both Sources. The practical application is that one user could control Source1 while the other user controls Source2.

To interface the USER port to a PC, a suitable USB to 3.3V TTL adapter is required. There are a number of sources for USB to TTL cables and adapter modules. Some soldering will be required in order to interface to Hirose User Port connector.

An example of a USB to TTL adapter is the FTDI UM232H from Digi-Key. There are simpler adapters and cables available from other suppliers. If you need help or want more specifics, just e-mail or call us.

User Port Pin Assignment Table
The USER PORT mating connector is a Hirose DF11-8DS-2C with pre-terminated cables H3BBT-10112-W4. All available from Digi-Key or you can order the 5009 List Mode Selector switch  (LMS-1).

The Valon 5009 has some interesting and novel features. For example, there are two serial ports available. One, micro-USB, the other a 3V TTL. You can use them both at the same time. Since it is a dual synthesizer, you could use the USB serial port to control the Source 1 frequency and use the TTL port to control Source 2 frequency.

The 3V TTL serial port uses a small Hirose connector and was intended to allow users to control the synthesizer in an embedded application directly from a local microcontroller. The 3VTTL serial port shares the User Port with the List Mode control.

Contact us if you have any questions..
________________________
 
Network USB Adapters
We have checked out the Belkin F5L009 with the Valon 5009 synthesizer and had no problems running either the GUI or the terminal program over the LAN/USB connection.  If you have discovered other adapters that work for you, please let us know.