[dorkbotpdx-blabber] Fwd: openSpectrometer PCB help and pattern matching software

[Nathan McCorkle]

So there's this contest/challenge:
http://publiclaboratory.org/wiki/spectral-challenge#challenges

Generally I'm not too interested in contests/challenges, but maybe
some of you are.

My openSpectrometer project is a bit stalled right now, PCB layout
isn't that interesting to me so learning and playing with the software
isn't the first thing I do for fun... I've got the schematic done up
in KiCAD

http://openspectrometer.com/


Also I've been working a lot with Artificial Neural Networks, and I'm
thinking it might be fun to hack on some spectrum matching programs
using them.

Anyone interested?

--
-Nathan


--
-Nathan
_______________________________________________
dorkbotpdx-blabber mailing list
dorkbotpd...@dorkbot.org
http://music.columbia.edu/mailman/listinfo/dorkbotpdx-blabber

[Nathan McCorkle]

Ahh, you're right, I don't have the KiCAD files up online. I wanted to
test it before making it public, which I haven't done yet because I
haven't got a PCB fabbed (from OSHPark of course!)

On Tue, May 14, 2013 at 2:58 PM, Erik Walthinsen <om...@omegacs.net> wrote:

> On 05/14/2013 02:49 PM, Nathan McCorkle wrote:
>>
>> My openSpectrometer project is a bit stalled right now, PCB layout
>> isn't that interesting to me so learning and playing with the software
>> isn't the first thing I do for fun... I've got the schematic done up
>> in KiCAD
>>
>> http://openspectrometer.com/
>
>

> I don't see anything from Kicad, but I might be interested in helping out to
> give my poor brain a break from work. I use Eagle though, so don't know if
> that'd be an issue.

[Jerry Biehler]

Neat, I sold my spectrophotometer a while back when I needed moolah. I have a SPEX .22m scanning monochromator spectrometer that I want to get running in it's stead. It has a stepper in it that moves the grating inside and takes step/direction signals. It has no sensor but I have a few Hamamatsu Integrated PMT blocks I can use. I was thinking of using one of the new teensys with their higher res analog in.

The unit is like this but without the adjustable gratings. http://www.ebay.com/itm/Spex-1681-0-22m-Wavelength-nm-Spectrometer-1681B-/370801706386?pt=LH_DefaultDomain_0&hash=item5655810592

Big problem for me is going to be software. 

-Jerry

[Nathan McCorkle]

very cool, I'd love to help setup a PMT. I haven't seen the high-res
teensy analog in, but I'm working on this analog design for the CCD so
maybe it would be useful for a PMT, I could imagine the PMT outputting
a digital frequency I guess.

[Jerry Biehler]

A PMT is a purely analog device. You supply power and you get a signal out proportional to the amount of light coming in to the unit. There is an external programming wires that takes an analog voltage to set the voltage level of the tube. This affects the sensitivity of the tube. 

This is why I was thinking of building it around one of the new teensy 3.0 boards. With the 13bit analog in it would give more than enough resolution. Pretty simple concept, I quick scan to find the peak signal and adjust the gain of the pmt to get something like 90-95% full scale and then go back and collect data points for each step of the motor. Calibration would be made with a standard He-Ar glow lamp. 

The step-direction interface with home switches seems to be pretty standard, this could be used to make a lot of old spectrometers or scanning monochromators useful again. 

The modules I have are like these but the window is on the large flat and not the edge. 

http://209.73.52.252/assets/pdf/parts_H/H9306.pdf

-Jerry

[Nathan McCorkle]

Hmm, some PMTs I've read about send out ticks, one per photon count.

[Jerry Biehler]

Yes, they will when used in something like a scintillator for gamma spectroscopy. But in a situation with a spectrometer there is so much more light it will be an analog signal.

-Jerry