Wonderfull  project . 

   Last year I run across old   sci.article: http://www.researchgate.net/publication/227816958_Urine_analysis_by_laser_Raman_spectroscopy

showing how Raman  can measure   molecule  in Urine (for my wife health needs) ...  crazy tech ...  then searching for DIY  Raman I run on your project at forum where people are thinkering how to build one cheap  : 

http://www.sciencemadness.org/talk/viewthread.php?tid=23422&page=4

They mentioned your great project but someone stated that  you didn't provide  measurments  of some  compund  spectra.

so few questions :
1)  Can you please give feedback on their comment ?
2) I assume you compared it to some    commercial  device ... If yes , can you give some information?
3) Do you consider adding option for  Surface-enhanced_Raman_spectroscopy, might be needed  for my needs ? 

  If i can measure some molecules  in urine   I will  make  time  for the replication this summer assuming that the parts price is  roughly  600-1000$  ( as you said in the

http://publiclab.org/notes/flatCat/08-29-2014/ramanpi-the-3d-printable-raspberry-pi-raman-spectrometer  ...  lucikly 

3Dprinter  and hackerspace is in the neighbourhood )

 Btw today  I was at ki-cad workshop  ( at local hackerspace)  and there i heard about cheeper alternatives to OSHPark,  presentor was very happy with quality  of :  SeedStudio , Hackvana PCB ,  Smart protoyping, Dirt Cheap Dirty Boards   (free shipping, 15$ for  10 pcbs  )

All the best ,

p.s.   "24hours is never enough"  totaly with you on this one .

Hi..  I haven't seen the discussion on sciencemadness... I should start by saying, ramanPi is still under development.  I started the project in late last year...  It's quite an undertaking for one person :)  So, a lot of progress has been made...but there is work to still be done.

I am in the process of trying to get as many developers together as possibly to help bring everything together.. I'm working on getting a FAQ and list of work to be done on ramanPi.org so people can see what needs to be done still and where they can contribute...  

So, to answer the questions...

1- Assuming the comment is regarding that I didn't provide measurements of any compounds, etc..  Long story short as to what has been the limiting factor.... CCD integration times due to processor limitations. The stm32f401re I chose originally for the imagingBoard doesn't have quite enough power to pull off what I needed..  I tried many approaches and nothing quite did it.. That is changing very, very soon..  Without writing a book here, I decided to go with FPGA for the imagingBoard..  This will accomplish two things..It will make the spectrometer portion more capable of the longer integration times, and will allow for extra signal processing, etc... And the major other benefit will be that it will allow the spectrometer to fulfill my original goal of making it work as a completely independent spectrometer outside the raman system if someone wanted just a spectrometer.. It will be fully compatible with the Ocean Optics USB4000..  What this means is that the end goal of obtaining actual raman spectra is almost here.  There's a lot of posts in the link you gave, but I hope this helps.. If not, let me know and I can clarify... :)

2- I have no access to any commercial systems (which is why I built ramanPi) =) But I am working with a couple people who do have access, and hopefully will be able to provide something there soon as well.. What I can say is that the resolution will be relatively comparable. (depending on the options one chooses to build)

3- I currently have no plans for adding options on surface enhanced raman spectroscopy..  Possibly in the future, but for now... I will focus on completing the base system.. :)  Having said that, there are a number of options I have planned for the future, so this may become one as well.. but if I recall about surface enhanced, it might be a little tough to implement on a DIY level...

It is my goal to bring the kits and parts to where someone can build a system for the $600-$1000 range...depending on options, etc...

Hope this helps!

"CCD integration times due to processor limitations. The stm32f401re I chose originally for the imagingBoard doesn't have quite enough power to pull off what I needed.."

I don't understand this.

HI Marvin..  I was driving the CCD clocks straight from the MCU.. I started with PWM, which I couldn't get to go more than about 500kHz.. tried using busout which was worse.  If I'm missing something, I'd love to hear it.. :)  I didn't add external hardware to drive the clocks mainly due to time constraints..  The FPGA redesign is already underway, thanks to some kind help.. 

You can toggle a GPIO pin every 2 clock cycles with the ST20F4 series so a (say) 1MHz read out even with math on the fly ought to be trivial on an 80MHz part.  Clocks up to 5MHz should be fine but much faster will start to smear the CCD and I'm not sure any application would benifit from the few ms savings in readout.  For integration, you should be reading out as few times as possible from the CCD until a bit below saturation.  For a Raman signal into the Toshiba CCD this could be 30secs while the MCU has no driving to do.  A slower chip could read into RAM and then compute the integration sums afterward but in this case you are probably ADC limited, 400ish ns if using the internal unit.  I've not looked at your design in some time, am I missing a part of your thought process?

I mean STM32F4 series of course, doh!

Marvin, sorry.. I think my memory failed..I think the PWM would actually allow around 1MHz..which I agree is still nowhere near what you'd think you could get. Correct me if I'm wrong but I believe it to be a limitation of the mBed libraries.. There is a 'FastPWM' library but if I recall it didn't support the F401 at that time.  I tried several variations including using the 'busout' function but the timing was worse..   I have had success reading from the CCD, but the trouble was reading the ADC while trying to maintain the master clock.. The documentation for the 1304 isn't great, but my understanding was that you need to drive the master clock during integration? while in the 'electric shutter' mode you read one pixel per four cycles..and in the normal mode SH stays low during integration but the master clock keeps going....?  When I mentioned the trouble with integration times due to processor limitations, I was able to readout using both the busout and PWM methods running at around 500kHz using the firmware that is in the gitHub...which uses the 'electric shutter' in a FSM .. It was when I tried moving to the normal mode and lengthening the SH period I ran into the problems..

I have read the ADC in the F401 is quite fast..and I do store the values.. I tried using the CMSIS RTOS library to see if that would help to read the ADC while driving the clock..but it was actually worse.  I haven't worked with any of the STMicro tools, which is probably the best way to go when trying this kind of stuff.. but in the way things have come out, the FPGA seemed like a much better way to go.....the desire to make the spectrometer a device that can be used on it's own, outside the raman system if people want to build just that...and be compatible with the ocean optics usb4000...doing a continuous acquisition while streaming the data compressed through the usb serial interface and the other various functions...it seemed that trying to make the F401 do all of that was too much to ask without spending a lot of extra time trying to be clever with coding..  With the efforts ramping up in development, and people starting to contribute...the FPGA looks like it's solving all of these problems..and hopefully won't add any addition cost..

I am more than open to any ideas you might have..or any info regarding the CCD.. :)

fl@C@ the Ocean Optics HR2000 uses a Xilinx XCR3128XL CPLD to drive the CCD , a LTC1415 ADC and a Cypress EZ-USB controller. Have you considered using PL chips instead MCU at the CCD data acquisition level? 

Hey, this is a fantastic job! When you will be decided to sell some kits, let me know, I want to buy one.

Hi..!! Will do..! They're in the works.. Hopefully in the following few months I will have some significant updates in that direction...!

Thanks!!

I got a lot notes on how to calculate the spectrograph. I got many free articles from archive.org, very good website.

Hey Fl@c@. THANK YOU SO MUCH for doing this great work.

one question, could you post the books or notes to calculate the czerny turner spectrometer? I 've been looking for a long time... thanks, or any online source or formula you can share with us?

Hey Fl@c@. THANK YOU SO MUCH for doing this complete build log. This is genius.

So, I WANT to build one really bad. I'm in need of this kind of technology right now and I have the money for the parts. I don't currently have a 3D printer though.

SO a couple of questions for you! There are 3D printer on demand sites on the web I'm sure. Do you recommend any? Never hunted for a service before. And you mention the PCB kits. You have any plans to sell them in the near future or should I just outsource some vendor who does Eagle? Can you recommend any prototype manufacturers?

Finally, your Eagle files are up to date? No dead bug mistakes or anything?

Hey FrankenPC.... So, thank you..! I am in the process of gearing up to offer kits.. I am planning on offering full kits, or parts.. I am also looking into offering either plastic parts (will probably be molded, or possible HDPE machined, etc..) and also aluminum parts.. The PCBs are current, but I will be updating them in the near future with more options... and I will be offering those as well. If you need something very soon, I've used OSHPark but they seem pricey to me..but the boards come out very nice. I don't know of any 3D Printer on demand sites...but I have seen a couple in shopping malls lately.. If you're not in a terrible rush, I am working very hard to get the site up and things in line to offer them... =)

Well you can count me as a purchaser when you do!
-James

I am a chemistry student at UIC and I have been working on some independent studies with spectroscopy, and I must say this is absolutely amazing. I was thinking of doing a basic visible light spectroscopy device with the raspberry pie, but now that I have seen this I may have found a new project to play around with. Please keep up the innovation!

Hi eichlerdr, thank you..! Maybe start out with just the spectrometer portion! Keep my updated on your build!

Hi bootdsc..! Hey, thank you.. I am just glad to have been a part of the whole thing, there were a lot of great projects.. I hope I can get to that point soon!

The filters are not symmetrical around 532. The short WL filter cuts up to 360 cm-1 from the laser line, this is good. The Stokes filter, otoh, cuts up to 615 cm-1, this limits the "good" side (most intense) of the scattering to higher vibrational frequencies only. Did you consider using a second gratting instead of filters to remove the laser exciting line?