Wyniki wyszukiwania - BazTech

1

Wpływ miejsca lokowania bramek na częstotliwość oscylatora pierścieniowego

Michalak S.

Pomiary Automatyka Kontrola

|

2014

|

R. 60, nr 7

444--446

PL

W artykule przedstawiono wyniki analizy zmian częstotliwości oscylatora pierścieniowego, uzyskane poprzez zmianę miejsca lokowania pojedynczych bramek w różnych częściach struktury reprogramowalnej. Przeprowadzono symulacje komputerowe oraz wykonano badania eksperymentalne układu oscylatora pierścieniowego złożonego z N inwerterów, dla kilku wariantów rozmieszczenia bramek i wykorzystania linii połączeniowych pomiędzy obszarami CLBs. Układ oscylatora implementowano w strukturze FPGA (Spartan-3).

EN

This paper presents the results of investigations how the inverter location in the area of a reconfigurable FPGA chip influence the properties of a ring oscillator. Ring oscillators are very often implemented in FPGA structures, even in the very advanced projects. They are used both as a single element or an array of sensors for measuring the chip temperature and thermal verification on reconfigurable systems [1, 2], as well as for measuring the propagation delay on the internal wires of the FPGA chip [3, 4]. In our investigation the ring oscillator composed of 11 inverters was implemented in the Spartan-3 structure (Fig. 1). There were performed simulations and experiments. We tested whether and how the location of the single inverter and the delay of lines influenced the ring oscillator frequency (Figs. 2 and 3). The properties of different connections between CLBs in the FPGA structure are described (Figs. 4 and 5). The ring oscillator was located in different areas of the chip to minimize or specially increase the length of lines between the inverters (Figs. 6, 7 and 8). The simulation and experiment results are presented in Tab. 1 and discussed. In conclusion we can state that when one wants to use a ring oscillator as a sensor and to analyze the frequency or delay times, there should be considered not only the influence of temperature or voltage supply of the chip core [8] but also the location of the sensor. In the case of an array of sensors, each ring oscillator should be analyzed and calibrated independently.

2

Ocena efektu samopodgrzewania układu PLD w niskich temperaturach

Arnold K., Pająkowski J., Michalak S.

Pomiary Automatyka Kontrola

|

2012

|

R. 58, nr 7

641-643

PL

W pracy opisano zachowanie programowalnego układu PLD po poddaniu go działaniu niskich temperatur, obniżanych do temperatury ciekłego azotu. Przedstawiono wyniki badań eksperymentalnych, zmierzających do określenia wpływu temperatury i poboru mocy przez wykonaną w technologii CMOS strukturę EE PLD na czasy propagacji zintegrowanych bramek logicznych. Zaprezentowano charakterystyki średniego czasu propagacji pojedynczej bramki w zakresie niskich temperatur i porównano uzyskane wyniki z prognozami, formułowanymi w oparciu o zjawisko samopodgrzewania struktury półprzewodnikowej.

EN

In this paper behavior of a programmable logic device (PLD) in the low temperature range, including temperature of liquid nitrogen, is presented. There are given the results of experiments in which we tried to determine the influence of temperature and power consumption on the propagation delay of integrated logic gates implemented in an EE PLD CMOS structure. The thermal conditions of work resulting from the ambient temperature, clock signal frequency, value of voltage supply and current consumption connected with output loads and switching frequency are discussed. The PLD device properties in the nominal range of ambient temperatures and expected behavior after reducing the temperature are described. The main idea of the circuit for average propagation delay measuring (Fig. 1) and the voltage-current dependence for recommended test output loads (Figs. 2 and 3) are discussed. The test circuit with pull-up resistors for increasing self-heating effect is proposed (Fig. 4). The results for the propagation delay (Fig. 5) and current consumption (Fig. 6) at 1 kHz and 1 MHz switching fre-quency as a function of the temperature changing from -196°C to 20°C are shown. The propagation delay vs. temperature (Figs. 7 and 8) and the current consumption vs. temperature (Fig. 9) for the circuit with external pull-up resistors are presented. The influence of voltage supply value changes on the obtained results is taken into consideration. The results are discussed and compared with expectations.

3

Badanie właściwości oscylatora pierścieniowego w temperaturze 77 K

Michalak S.

Pomiary Automatyka Kontrola

|

2012

|

R. 58, nr 7

681-683

PL

W artykule przedstawiono wyniki eksperymentów, w których testowano działanie oscylatora pierścieniowego zaimplementowanego w układach reprogramowalnych. Analizowano właściwości opóźniające inwerterów zaprogramowanych w strukturze CPLD układów XC2C32 (Xilinx). W temperaturze otoczenia (300 K) i w temperaturze ciekłego azotu (77 K), badano zdolność do generacji drgań, stałość częstotliwości oscylatora (na podstawie pomiarów średniookresowych), wpływ zmian napięcia zasilania na częstotliwość oscylacji.

EN

In this paper the results of experiments with a ring oscillator implemented in programmable devices (XC2C32 Xilinx) are presented. The examined devices were immersed in a Dewar flask (Fig. 1) with liquid nitrogen. It was found out that the ring oscillator (composed of 11 gates) (Fig. 2) still worked properly in such low temperature. According to the theory of silicon semiconductors, the activity of carriers increases in low temperatures, so there was expected decrease in the propagation delay for every gate and increase in the oscillation frequency. The output frequency was measured and the average propagation time for inverters was calculated. The results at 77 K (temperature of liquid nitrogen) were compared with those at 300 K (room temperature) (Tab. 1). The output frequency characteristics versus the supply voltage for the examined devices were measured and drawn (Figs. 3 and 4). The quadric polynominal functions which fit these non-linear characteristics were proposed. The relative change of the oscillation frequency versus the supply voltage is shown in Fig. 5. The frequency sensitivity depends both on supply voltage and temperature. The relative sensitivity (normalized) in relation to the voltage at 300 K and 77 K is presented in Fig. 6. Based on the results from 24-hour measurements (86400 samples were collected) the frequency stability was determined. The average value and standard deviation value were calculated (Tab. 2) but first and foremost there was calculated and plotted the Allan deviation (Fig. 7).

4

Pomiar czasu propagacji inwerterów implementowanych w układach programowalnych w temperaturze ciekłego azotu

Michalak S.

Pomiary Automatyka Kontrola

|

2011

|

R. 57, nr 8

828-829

PL

W artykule przedstawiono wyniki eksperymentów, w których sprawdzono zdolność pracy wybranych układów reprogramowalnych w warunkach niskich temperatur, w szczególności w temperaturze ciekłego azotu 77 K (-196°C). Zaprezentowano wyniki pomiarów uśrednionego czasu propagacji inwerterów zrealizowanych w testowanych układach reprogramowal-nych. Do badań wybrano struktury PLD oraz CPLD. Badano układy ATF16V8 (Atmel), GAL16V8 (Lattice) i XC2C32 (Xilinx). Przedstawiono wyniki pomiarów uzyskane w temperaturze pokojowej (300 K) oraz w warunkach niskich temperatur - w temperaturze ciekłego azotu (77 K).

EN

In this paper the results of experiments with programmable devices in low temperatures are presented. For most CMOS devices, including programmable devices, low temperature, in particular the temperature of liquid nitrogen 77 K, is far below the typical range. The producers usually guarantee the proper work for their devices at 0°C for commercial devices or -40°C for industrial ones. Even for special military devices the lowest temperature used is -55°C. In the experiments performed by the authors the ability of proper working at the liquid nitrogen temperature (77 K) for some chosen PLD and CPLD devices were examined. The examined devices were immersed in a Dewar flask (Fig. 1). There was found that some of them worked properly in such low temperature, and also could be programmed. According to the theory of silicon semiconductors, in low temperatures the activity of carriers increases, so decrease in the propagation delay of the measured gates was expected. There was measured the average propagation time of the inverters implemented in programmable devices (Fig. 2 and Fig. 3). The results for GAL16V8, ATF16V8 and XC2C32 are given in Tabs. 1 and 3. The obtained results of the average propagation delay and the estimated maximum system frequency were compared with the datasheet information (Tabs. 2 and 4).