Tuesday, 4 August 2015

Nanosecond pulse generator

Well, where to start... A nanosecond pulse generator is used as part of an experimental space charge measurement equipment developed at the UNSW's High Voltage Laboratory. The required output must provide not only a narrow width (less than 15 ns) but also a magnitude that exceeds 200 V (the goal was 500V). Most of the literature available regarding nanosecond pulse generators is related to biomedical applications, e.g. Sanders et. al, [1] have developed a pulse generator to deliver electric fields to biological loads in order to produce cellular membrane  electropermeabilization. Pulse generators of this type (relatively high magnitude, short duration) are being used in other applications within four major categories [2]: 

  • Industrial: food processing, concrete recycling, plasma systems.
  • Environmental: ozone generation, waste water treatment.
  • Medical: electroporation, plasma medicine.
  • Military: laser guns, electromagnetic launchers, radars.

Nanosecond pulse generators are available in the market but they are yet expensive, especially for research and educational activities. The cost of a single pulse generator with the required characteristics can go well beyond 4000 USD. AV Tech pulse, FastPulse technology Incorporated, Yamabishi Corporation, FID GmbH are among the companies specialized in providing this type of equipment. 

Some of the most important topologies are:
  1. Diode opening switch (D.O.S): this topology generates a clear and well defined pulse, however the calculation of the components is complex as well as its implementation since it depends on several parameters such as the reverse recovery time of the diode, the MOSFET’s linear behaviour and the spurious triggering in the PCB layout [1].
  2. Marx Bank based pulse generators: They are usually big in size and their implementation require spark gaps that may lead to high jitter. There is a variation of the topology called miniaturized Marx Bank which use transistors instead of spark gaps [3].
  3. Transmission line based pulse generators: self-matched transmission line Blumlein [4] provides fast rise time and a square shaped pulse [5].

The first topology had been chosen due to its simple circuitry and was successfully implemented by Wu Chao [6] based on the work exposed in [1]. My first task was to make copies of the existing pulse generator but changing the pulse width in order to see if the pulse width has a significant impact on the measurement results. Unfortunately, Wu Chao had already graduated by the time I was doing this project so I had no have chance to bother him with questions. After reading his work, I decided to simulate the circuit using LTSpice, this was crucial for me to understand how this topology works so I think it is worth to share it. It does not have the same components used at the end but it gives a good approximation of the result. The main component is the diode D1 since the pulse width and the selection of C1,C2,L1 and L2 values depends on the reverse recovery time of the diode [1].

Simulation circuit of the pulse generator using LT Spice. Stop time=0.1s.

Simulation result of the pulse generator circuit using LTSpice. The pulse output can be noticed in blue in the bottom panel.
After few weeks, a couple of pulse generator boards were made, they are capable to provide a pulse width of 13 ns (full maximum half width method) with a magnitude of 250 V.

Pulse generator boards. Courtesy of  UNSW's High Voltage Laboratory.

Pulse output. Courtesy of  UNSW's High Voltage Laboratory.

References
[1] J. M. Sanders, A. Kuthi, W. Yu-Hsuan, P. T. Vernier and M. A. Gundersen, “A linear, single-stage,nanosecond pulse generator for delivering intense electric fields to biological loads,” IEEE Transactions on Dielectrics and Electrical Insulation, vol. 16, pp. 1048-1054, 2009.
[2] S. Zabihi, Flexible high voltage pulsed power supply for plasma applications, Brisbane: Queensland University of Technology , 2011. 
[3] M. Inokuchi, M. Akiyama, T. Sakugawa, H. Akiyama and T. Ueno, “Development of Miniature Marx Generator Using BJT,” IEEE Pulsed Power Conference, pp. 57-60, 2009. 
[4] S. Romeo, C. D’Avino, O. Zeni and L. Zeni, “A Blumlein-type, Nanosecond Pulse Generator with Interchangeable Transmission Lines for Bioelectrical Applications,” Transactions onDielectrics and Electrical Insulation, vol. 20, no. 4, pp. 1224-1230, 2013.
[5] G. Peng, Design of a modular high voltage nanosecond pulse generation system, Sydney: The University of New South Wales, 2013.
[6] C. Wu, Space Charge Measurement in Solid Dielectrics, Sydney: UNSW, 2012. 

Sunday, 2 August 2015

USB atmel programmer

Back in time, when I was an undegrad student, I used to have an Atmel programmer with a parallel connector. Happiness was endless burning micro controllers one time and another, until I upgraded my old but loyal desktop PC for a fancy laptop. I won't lie, I get used to the laptop rapidly with not even a minor remorse about leaving my PC aside. However, the laptop did not have a parallel port. But it was plenty of USB ports, so my first option was to get a parallel-to-USB adapter but just before that happened I lent my programmer to a friend and that fellow has not show up till now. Months later, I looked for a way to make my own programmer to satisfy my thirst of doing electronics and got to the website of Thomas Fischl and his USBasp (http://www.fischl.de/usbasp/) which was simple and cheap. Within a few days I got the components, spent some time doing the routing in Eagle and voilá! a USB programmer was ready to be used... well it was not that easy, the USBasp uses a ATMega8 to deal with the communication protocol between the computer and the device to be programmed. So basically I was in need to use another programmer to burn the code into the ATMega8. Happily for me, there were some spare programmers at my uni that I can borrow. Once that was complete the rest was only the installation of the firmware and a software(eXtreme burner) to write/read into the microcontroller which I also got from Fischl's website.

There are a couple of jumpers in the circuit proposed by Fischl. they allow clock selection (JP1) and self programming (JP2). In my case, JP1 is connected to ground and JP2 had left open since I was not using those features.

Here is my programmer, it is more than 3 years old and it has been used in Peru and now is helping me to do some tasks during my masters studies in Australia. I hope you guys can also enjoy your free time doing some cool stuff using this programmer. If you feel a bit lazy or wanna go safe, the USB programmer can be purchased from Thomas' website.