High speed camera for analysing motion control

[Plc_User]

I would like to invest in high speed camera that I would use t analyse problems in high speed processes.
f.i. machines with fast motion control, where the changes are too fast to observe.
I mean a camera like a commercial camera to take pictures and record movies, but that is fast enough to capture high speed processes, and that afterwards can be used to replay the movie in slow motion.
Anyone experiences or advices?

 

[dlweber]

I know there are ones built especially for what you are talking about, stronger design, not the highest resolutions to save cost but many industrial needs don't need top notch picture quality or 18 billions colors,

Space limitations are part to if you want to run these cameras continuously you are probably looking at 1G ever 4-6 seconds (glanced at a few models online) if you went with HD cameras, well we will just say that will take ALOT of disk space

 

[rdrast]

How fast do you want?
Consumer level cameras like the Sony NEX FS700R can do 960 frame per second bursts for about 15 seconds at around $6000.
The Photron UX50 Mini can do up to 160,000 FPS, but for practical video is limited to about 5000 FPS, and retails for about $48,000.
The Photron Fastcam SA-Z can handle up to 2,100,000 FPS, but for practical video is about 75,000 FPS, and retails for well over $100 K.

 

[danw]

Discussion here generated by a relatively inexpensive camera,$400 (5years ago) that does 1000 frames/second in burst mode.
http://www.eng-tips.com/viewthread.cfm?qid=246142

 

[Peter Nachtwey]

a decent motion controller should be able to log positions, velocities, status bits, error bits every scan for a few minutes. A high speed camera shouldn't be necessary for most applications.

We have an Olympus iSpeed2 that can record up to 38K FPS but the lighting must be good. In practice 2K FPS is more likely even with halogen lamps. our motion controller can log 16 item at 4K for a few minutes and freeze the data when a trigger occurs so that what happened before and after the event is captured. this is good for those glitches that occur only once in a day or week.

 

[kamenges]

rdrast is correct that you first need to define how fast this thing needs to be. I suppose in practice you could set your budget and then get the fastest thing you can afford but that seems backward to me.

I think 1000 FPS would cover most of the things most of us face on a daily basis. I'm sure their are exceptions but that would get you that capability to detect an oscillation approaching 500 Hz and see incremental motion of 0.2" at 1000 feet/min.

I partly disagree with Peter's assertion about the relative merit of high speed video compared to motion controller data logging. They are for two different things. Peter deals with what most of us would consider tightly coupled, ridged systems, although he often doesn't feel that way. He can reasonably expect that the business end of what he is controlling is positively linked to his actuator and that the actuator position is an accurate measure of the load trough some mechanical transform. I that case motion controller logging is extremely valuable. However, in my previous life I worked on paper sheeters. We had a vacuum modulation assembly that pulled the back end of a sheet down so the leading edge of the following sheet could pass over it. High speed video was invaluable to us in seeing how the leading edge of one sheet was interacting with the trailing edge of another. In this case the linking between the actuator and the actuated was free air.

Keith

 

[Peter Nachtwey]

They are for two different things.

Yes.

About the high speed cameras.
I have found that there is a practical reason/limit as to why most cheap high speed cameras limit the FPS to 240. Recording at 1000 FPS will require more lighting.

I like to play table tennis. I play at the local club and at work. Table tennis players are mostly uneducated and have built up a mythology and terminology about the game. I have debunked a myth or two by recording the impact of a table tennis ball with my paddles at 2000 FPS. I can tell you that standing in front of 2500 W of halogen lighting gets hot and it is hard to see the ball staring into the bright lights.

 

[shooter]

It is a lot easier to capture an analog value from a potmeter or alike, then using a camera, as you will need to interprete the whole.

 

[CapinWinky]

I would highly suggest that you get a camera that does not require a mini sun for lighting. I used a very fancy high-speed camera to run around 1000 FPS and couldn't get a good image because I couldn't get the camera and the halogen lighting into the space I needed to film. I ended up using a GoPro, which isn't very fast.

I would suggest a high end consumer camera that offers high speed video, they usually work very well with low lighting. I know Nikon 1 (search Nikon mirrorless) offer decent high speed and can do it with standard lighting. The higher end Sony NEX cameras also have high speed, but the lower end ones can't go as fast. Download the manuals and double check the max FPS before buying.

As a bonus, a good consumer camera can also be used for things other than high-speed video.