LENRCAM?
We've had a bit of fun dreaming up a possible way to SEE LENR on live video footage. Follow along and I'll rationalize. . .
If a nuclear reaction is happening inside these cells, there's got to be a lot of thermal energy. But we're talking about, dare I say it, COLD fusion here and we search for much smaller amounts of excess energy. Maybe the anomalous heat we're all looking for is a macroassebmly of nanoreactions. These localized areas would produce extreme heat commensurate with a nuclear reaction and assemble into controllable and measurable excess heat.
In v. 10 of JCMNS, David J. Nagel opened the issue with the possible correlation between materials that have exhibited excess heat and craters that have developed on the surface. Though Nagel could not make any solid conclusions for or against this proposition, it got us to thinking. We can't comment if this idea is valid or not, but we think we have a way to get us going in the right direction.
It's not as difficult as it may sound, and it actually doesn't involve a thermal camera. Thermal cameras are costly and can't even see through glass, making the prospect of looking at the metal surface in our reactors impossible. That's where all the action is and that's exactly what we CAN'T see (thermally, of course).
Turns out that Paul knows a thing or two about black body radiation, and also knows a lot of things about electronic devices. So we grabbed a webcam and got to a google hangout not to see people, to see infrared radiation.
However as cool as this sounds we have no way of processing the resulting images to store the events the camera would witness. We currently believe saving the histograms from each pixel of the megapixel camera would be the best way, but that would eat 1GB+ of RAM, and be a lot to sort through. We need help from you AV gurus. What do you think? Do you have any suggestions for this dilemma?
Comments
views. I know this is completely off topic but I had to share it with someone!
When we had moved the cell out from under the hood last week it did get to room temp. It did not however go to room temp when we had the bug. It temporarily went down to half power and then back up.
Yeah that's amazing about assembly. I use turbo pascal and just like c++ the conversion to assembly is minimal. Also I make sure to directly access all the bitmap data with pointers. Id be really surprised if assembly can make this twice as fast without using the GPU. I did a speedtest comparison to java once, some high level things like making a billion nodes in a tree goes 20 times faster in Turbopascal (In Delphi) than it does in Java. So much overhead in those languages.
Some years ago a made a program that ran 100 robots on a playground. The ASM algorithm was about 100 times faster then in the original language. It was fun to do that. A pitty that ASM gets more and more forgotten.
Your Program seems to run fast enough though
@MFMP: If you need some help with VBA or something let me know.
I Made an initial test and it can do a 640 by 400 pixel bitmap at like 1000 fps+ on just 1 thread of my 2.8ghz CPU which is like 20 times faster than the webcam itself. Which can even be ten-folded by running it on the GPU.
Thanks for the offer though, assembly people are hard to find these days, its a dying but beautiful art.
I send Paul a demo, we'll see where it goes from here, if we want to make this stream-able online a lot of work still has to be done.
IF there is the need to do the core in Win32 ASM because it runs too slow I could have a try. (only if it is just a calculation with two pictures at once like: are there changes I need to highlite in one of the pictures) It is just a problem with my timeresources at the moment.
Yes, the history worked for me. I can see the data and I was surprised to see that the Pxs has grown so much!
When the power went off before, did the temp of the cell drop all the way to ambient?
Contact me at paulhunt
We are seeing a delay in the server to catch up to the averages data type. They should be up in a day or so. Until then you can see the data by changing the type from 30 second averages to History for now.
I at least still cannot see any data, I'm using this address data.hugnetlab.com/, It's obviously active but maybe you're not streaming the data here anymore?
Yes logging it will be the easy part, i could make it post statistics online or to a file.
What will be a tad harder is to livestream it but can also be possible. I have set up a bit of a concept for you guys to check out. Is there a way for me to get in contact besides this website?
You have a good idea.
I am also looking for ways to gather data about how many spikes, where, and of what intensity over a period of hours.
I have a few concepts. Do you have any ideas?
It should be up :)
I imagine we will be searching for tiny little spikes in the video footage
What I would do is every frame check every pixel against the same pixel last frame, if it goes lots brighter we have a little heat pulse. Also the frames will cancel eachother out so you will only see a black image with spots that represent change.
I hope this is clear, email me if you want are interested :) (Or if you can't see the email leave a comment.)
opencv.willowgarage.com/wiki/
Time stamp both.
I would recommend using a Hero Black - as it has the ability to do frame rates from 1 in 60 seconds to 240fps - built in H264 encoding and uncompressed HDMI out - use of which with a capture card could allow CV data processing without data loss. But not that cheap!
gopro.com/.../hero3-faqs
the first problem with IR pictures is, that you need very expensive lenses made of quartz or you try a mirror optic which is fully transparent for IR due to the construction.
If you want to go this way, I would use a mirror system with short F and a BW CCD camera made for astronomy.
theimagingsource.com/.../...
These special cameras have a better sensitivity as webcams and the spectrum covers a part of the IR. They are available in a price range, starting from abt. 300$ and the software to grab the video is mostly free of charge.
RSS feed for comments to this post