Tuesday, October 27, 2009

N2PK VNA - USB to Parallel interface

I am pretty happy with the Parallel port interface - it doesn't require any drivers and works just fine. For portable applications using my Netbook in the field, I have no choice but to use an USB interface. My Acer Netbook doesn't have PCMCIA slot (for Parallel card) or native parallel interface and the only option to control the VNA is USB. A standard, off-the-shelf USB to Parallel port converter (commonly sold for printer interfacing) won't work. These converters are designed for printers use only and they don't have 100% bi-directional parallel port support. Fortunately, Dave G8KBB designed a very nice interface based on the Cypress FX2 (EZ-USB) USB micro-controller (CY7C68013A).
For my project I used a PCB from WB6DHW. His interface is a nearly identical clone (electrically) of the G8KBB interface. To be honest, I am less than impressed (other words, which I'll save come to my mind ) with the PCB layout done by WB6DHW - it looked like somebody was learning how to design a PCB and used this project as a practice board. I was almost ready to design my own board (G8KBB's board is excellent but his PCBs are not readily available). Don't mean to bash WB6DHW - just expressing an opinion. At the end of the day tho, I decided to close my eyes and got the bare board from WB6DHW because of its low price - fabricating my own board was going to cost a lot more and wasn't worth the effort/money just for a single piece. All components are from DigiKey, including the Cypress chip and the Hammond die-cast aluminum enclosure housing the interface.


I had to modify (mill) the PCB in order to fit it in the enclosure. The PCB is raised on stand-offs since the Mini-B USB connector along with some other parts are placed on the bottom side. I prefer to have the much sturdier and more reliable USB-B connector but there wasn't an easy way to modify this board. The Cypress FX2 chip is in a package with fine pitch leads but it wasn't that difficult to solder it pin-by-pin under sufficient magnification. There was no need to install any of the other connectors - I just soldered the wires directly to the board. The enclosure is a tight fit and if I had the connectors installed it would have been difficult to manage the wires inside. The most time-consuming part in this project is wiring the board to the cable and the female DB25 connector according to the schematics and the table provided by Dave G8KBB.

This is the complete interface. It is rugged yet compact - the only delicate part to worry about is the Mini-B connector. The initial setup is a bit complicated - a Vendor ID and Product ID must be written in the EEPROM in order for the USB micro-controller to properly report the interface in Windows. To accomplish this task, a step-by-step procedure and a piece of configuration software are published on Dave's, G8KBB web site. First, special drivers are installed in Windows to get access to the Cypress FX2 chip and the EEPROM address space. Then, using "USB Configure" (by G8KBB), the appropriate IDs are programmed in the EEPROM according to the hardware in use - version of the interface/VNA and current demand (in case the USB port is used also to power the VNA). This setup is one time deal - afterwards the interface is used with its regular USB drivers.

1 comment:

Andrey E. Stoev said...

Update: There was a problem with the VNA calibration while using the USB interface, resulting in inconsistant measurments and weird plot shapes during scanning. After some troubleshooting, the problem was traced back to the "filtered" DB25 connector. I installed such "filtered" type connector because of the 12 ft length of my parallel cable. As it turned out, the 1300 pF capacitors inside the connector were affecting the timing/waveform of the USB i/f pulses. For some reason this affected only the calibration routine (possibly the software was accessing the VNA at higher speeds during calibration). After replacing the connector with standard DB25, the problem was solved. There are RC filters on the control lines on the VNA board and they are working well - I did not notice any additional noise because of the long parallel cable.