A prototype is a preliminary model of something. Projects that offer physical products need to show backers documentation of a working prototype. This gallery features photos, videos, and other visual documentation that will give backers a sense of what’s been accomplished so far and what’s left to do. Though the development process can vary for each project, these are the stages we typically see:
Proof of Concept
Explorations that test ideas and functionality.
Demonstrates the functionality of the final product, but looks different.
Looks like the final product, but is not functional.
Appearance and function match the final product, but is made with different manufacturing methods.
Appearance, function, and manufacturing methods match the final product.
About this project
VNAs (Vector Network Analyzers) are a critical piece of radio frequency test equipment. And unfortunately, standalone VNAs are extraordinarily expensive. New VNAs typically cost at least $10,000 - with additional costs for calibration kits (manual cal kits: >$1000, automatic ~$10,000). However, VNAs are not doing tasks fundamentally different than what your smartphone does: transmitting a specified frequency, and reading back the magnitude and phase of a return signal. Frustrated by the high cost of this equipment, I decided to make a VNA myself. There is no software to install, no drivers to add, not even an ON button: plug it in and you're ready to go.
The Chazwazza is a 400MHz-2.7 GHz VNA capable of measuring S11 and S21, with display formats VSWR, dB Magnitude, and Smith Charts. Calibration is done with an onboard integrated calibration unit, and all units ship with the required power cable, SMA female ports and a semi-rigid coaxial cable.
What Can You Do With A VNA?
Most importantly: you can design, tune and validate antennas in the frequency ranges including GPS, GSM/3G/LTE, 2.4GHz WIFI, BT, and more. VNAs enable you to measure the isolation between two antennas, and do RF impedance matching. Transmission line losses, filter responses and LNA gain can be measured as well.
Frequency Range: 400 MHz - 2.7 GHz
Points Per Sweep: 231
Measurements: s11 and s21
Calibration: Integrated 2-Port Calibration
Formats: Smith Chart, VSWR, dB Magnitude
Markers: Up to 4
RF Connectors: SMA Female (2)
Port Impedance: 50 Ohms
Power Input: 5.5/2.1mm DC Jack Input 9V.
Battery Powered: Possible with input voltage between 6 and 11 Volts (not included).
Display: 480 by 320 Pixels, 252k color 3.5” TFT LCD display
Coaxial Cable: 12" SMA male to SMA male Semi-rigid cable .086"
This project didn't come together without a lot of iteration and struggle. The origins of the project go back to early 2016 when I attempted a scalar network analyzer with discrete components:
In the summer of 2016 I made a development board connected to an STM32F4 discovery board:
After the development board I drilled down on the RF sections and studied the directional couplers:
In November of 2016, the first standalone prototype was developed:
Proto2 was the first form factor development board, developed in February 2017:
March 2017 led to EVT1, with a larger display and support for S21 measurements:
Finally EVT2 was produced, with minor hardware bug fixes from EVT1, and we have the current design. For an in-depth discussion of the development (when the project was not yet ready), see this talk by Pete at the Hardware Developers Didactic Galactic in San Francisco.
Where’s the Funding Goal Come From?
The funding raised will be used to produce a batch of fully assembled 100 VNAs. This cost includes developing injection molding tools for the enclosure, which enable cost-effective volume production of plastic parts. We will do a final PCB iteration, including making 100 PCBs along with purchasing/populating the boards with the components. We need to order the production quantities of the displays, screws, power cables, coaxial cables, etc. Then we're off to the last step: final assembly, test and packout (FATP).
How Accurate Is the VNA?
No instrument is perfect. But rather than make long-winded statements with specs, let's compare the Chazwazza results with an extraordinarily expensive Agilent VNA (which is also imperfect).
Here is a 3dB attenuator swept from 400-2700MHz, both axis from 0 to -18 dB:
A 2.4GHz antenna is compared next:
The s11 results for a dual band cellular antenna (824-960MHz & 1710-2170MHz) is shown next:
Let's move on to S21. Here is the isolation between two antennas sharing a PCB, again measured in dB Magnitude:
Finally - this is a Vector Network Analyzer after all, where's the phase? For that I present Smith Chart comparisons for the same cellular antenna above, over two frequency ranges. Pardon the lack of interpolation between points, (I have yet to implemented that):
The same antenna's Smith Chart for the highband is now shown:
There you have it - can you design and tune antennas and RF systems with this product? Absolutely. There is some measurement noise that I'm working on filtering out with software, but overall the data is very accurate.
What type of OS does this run?
There is nothing to install: plug it in and go. This product uses a simple user interface that can be navigated via Next/Previous buttons, along with Enter to select, Increment/Decrement to adjust values, and Escape to enter or exit the menu from the measurement mode.
Up to four frequency markers can be selected. Currently the number of points per sweep is fixed at 231. I hope to add support for open circuit port extensions and memory traces, but this is not currently implemented.
Who’s Behind this Project
Peter Bevelacqua is a longtime professional antenna design engineer, having designed antennas for major companies including Boeing, Apple, and Google. He has many antenna related patents, a Ph.D. in antenna engineering and is the author of antenna-theory.com.
This Chazwazza is designed and manufactured in Silicon Valley, USA.
Risks and challenges
The good news is the project has existed for quite some time and is already on its 5th iteration, so most of the development kinks have been worked out. However, the engineering challenges of creating one functioning system is quite different than the challenges of creating 100 systems. There will likely be some hiccups and learnings in the production run, but given the successful level of execution thus far I am confident in being able to deliver.
There are some relatively minor software features that still need implemented, such as adding marker values on the screen, enabling memory traces on the display, and adding support for open circuit port extensions.