High-Sample Rate Oscilloscope or Galvanometer

[calbiterol]

I need to build / acquire (cheaply) a galvanometer or oscillioscope (preferrably oscillioscope) with a high sample rate with a USB connection. Any suggestions?

 

More in depth description:

I am doing an experiment that involves measuring the induced backspin in spherical projectiles. I have not (yet) calculated the velocity that the projectiles are travelling, but suffice to say the launcher packed enough punch in the low-power testing phase to put ~3" long dowel rods through half inch laminated high-density particle board.

 

I plan to measure backspin by embedding spherical plastic projectiles with NIB magnets. These will be pre-aligned before firing so as to rotate along the correct axis. Immediately after the acceleration tube (barrel) there will be another tube for backspin analysis. There will be a wire coil on the bottom of this tube as well as on the top. This will allow me to calculate both the backspin of the projectile and the forward velocity.**

 

The problem is that the projectile will be travelling EXTREMELY QUICKLY. This translates to a very, very brief time in the sensor tube, and on top of that, the rotation of the projectile could be very high - meaning that the galvanometer / oscilloscope would need to be extremely sensitive and have a very, very high sample rate (I would guess on the order of 1 microsecond). I need this to digitally output to my computer.

 

Any suggestions? I've mulled over both a USB input using a microcontroller and using sound-analysis software and the microphone jack.

 

 

 

**The magnets will be arranged in a ring around the projectile, so as to maintain both balance and ability to be detected. Each time either "end" of the projectile passes the top of the tube, a current will be induced (likewise with the bottom). The average of the current spikes on the top and on the bottom will provide me with the linear speed of the projectile. (I may be forgetting to divide by two here). This can then be used to find the backspin.

 

It occurs to me that my logic with that explanation is slightly flawed. However, backspin will still be easily measure with that setup - in a far more straightforward way - simply by relating the number of spikes with the time.

 

 

That's the (somewhat discordant) idea, anyway.

 

Thanks much.

[raivo]

One sample per 1 microsecond is not that high. For example common video AD converters use 20 or 30x higher sampling rates. Unfortunately i do not know of microcontroller that has such AD converter built in.

 

It may be that you actually do not need continuous sampling (and transfering everything to computer). Cycles with, say, 2048 samples may do the job. In that case you can use common serial port for computer connection which makes things a lot easyer. You also my not need direct graphics output as everything could be written to file that can be converted to visual representations afterwards.

 

Such system can be built using logic circuits to make AD to run standalone and write its output to FIFO or RAM. After complete sampling cycle microcontroller reads memory content and sends it to PC.

[Externet]

Hi.

I would attempt registering the spin using a storage oscilloscope. Borrow, rent, or... It will keep the waveform for analysis of which sampling approach is more convenient if you still want to involve a PC into it, or if the sensor needs modification to obtain a better amplitude/detail.

Miguel

[TestGuru]

I think you can try some sound card based oscilloscope shareware first, like this one at:

 

http://www.virtins.com

 

some of the sound cards supports 192 kHz sampling rate, but I am not sure whether it is suitable for your case or not.