Re: Could you recommand a good Real-Time control software?

Thank you very much for your help!

Our application is on noise control with sampling frequency of 5000 Hz, it has two anolog inputs and two anolog outputs. We need to control the input & output in real-time.

At present, we used it on DOS system with a desktop. But we want to run it at a Laptop under Windows system using PCMCIA bus Data Acquisization card. We tried real-time windows target toolbox of MatLab. It didn't work well. So we are wondering any better real-time software that we can use it with C,or C++.

Thanks a lot!

Herman Family wrote:

Xun Yu,

That is some sampling frequency. I'm not surprised that matlab didn't work well. That is a generalized software which has to worry about all sorts of interesting things, and is not optimized for that high a data gathering rate.

Assuming that each data point is two bytes, you are moving around 10000 bytes per second, or about 80 kbaud, which is faster than a lot of modems, and very rough on software.

I think you have a couple of choices here:

Your choices are limited because of the high sampling rate. Using the chip could allow you to hit higher speeds, as it would communicate less frequently with the pc.

Michael

Walter: I don't understand your last statement here. As a rule of thumb the sample rate should be at least an order of magnitude greater than the bandwidth of the output device. ( Shannon, Nyquest et al) The most important effect of a higher sampling rate is to reduce delays which can kill your phase margin.

As to this application, this is Hard Real Time, and XunYu should look into a dedicated DSP. A trial version of VisSim-embedded can be used to model the application and code can be loaded to a TI DSP such as the F2812 which will support a 400 KHz sample rate. Trial versions of software, and eval boards should get him moving in the right direction.

Just my opinions.. Jeff Lowe

Jeff,

I agree that you need to sample faster than the output device. That is why I asked, "What kind of an output device are you driving?" Xun states he is trying to control noise level. He is sampling the noise level. If he is producing a signal to actually attack the noisy waveform, the controller and its output must be working in the same frequency range. But if all he is doing is measuring noise level and using this to adjust an engine throttle, or whatever, then the noise signal must be averaged and is then used to control the engine. In that case the engine will need a controller with a one second scan, no more.

To repeat, What is the application?

Walter.