[DIYbio] Re: [dorkbotpdx-blabber] Open-Source Spectrometer

I used KiCad, did the analog stuff and some more by hand, then pulled
it into the http://www.freerouting.net/ autorouter for the rest.

I didn't want to go 2-layer as all the low-noise analog design stuff I
read seemed like enough reason to go 4-layer. I think I might break
the board into two, which would let me choose different number of
layers for each, if 4 are really needed.

I'm also not sure if I need the datasheet-recommended audio
transistor, since I didn't seem to noticed any loading effect when I
connected my o-scope with and without the transistor. The way I've
arranged the analog section, if I remove the transistor there's about
2 or 3mm between the Analog output and the ADC input... while I
believe the datasheet may be covering cases where the CCD is in a
hand-held barcode reader and several feet from the rest of the
circuitry. I could be wrong though, so its in there for now.

Like I said earlier, I'm not even using the second inner layer, just
top and bottom, with an unbroken ground plane (this is what I really
was going for, I didn't want to cut into the plane, as a lot of analog
design stuff recommends that over the tying two plane sections at the
ADC).

On Wed, Jun 4, 2014 at 1:52 PM, Jerry Biehler <jerry.biehler@gmail.com> wrote:
> That board does not look nearly complex enough to need 4 layers. How are you routing it?
>
> -Jerry
>
>> On Jun 4, 2014, at 1:40 AM, Nathan McCorkle <nmz787@gmail.com> wrote:
>>
>> So... ~2.5 years later and I just ordered a PCB for the openSpectrometer:
>> https://github.com/nmz787/open-spectrometer/blob/master/pcb-design/propeller_square/trace_layers.png
>>
>> After tons of reading about electronics, it still isn't as good as I
>> can do, but I think it's a good start to at least make sure the
>> general circuit works as expected/intended... then I can think about
>> noise reduction a little more. For example I'm only really using 3 of
>> the 4 layers of the PCB right now, routing power traces through the
>> 3rd layer or changing the layer stack with GND on both exteriors is
>> also an decent idea, but having signals inside makes debugging hard...
>> but I'm pretty happy with this for a 'first' pass.
>>
>> There seems to be a lot of room on the PCB, but I wanted to keep as
>> much digital out of the ground plane around the analog area. After
>> debugging I'll probably want to split the board into two, so I think
>> the room will get diced away and lead to slightly lower board cost
>> altogether.
>>
>> Right now, for this 4-layer board, 3 copies of the PCB from OSHPark
>> was $67.... so <$33 per PCB. Previously I estimated about $46 in
>> parts, so 46+33 = $79 with no optics (or SD card or USB cable or
>> housing or upgraded/coated CCD window).
>>
>> I've added the current files to github, but I'm not sure all the
>> libraries are there, but that shouldn't stop a determined person from
>> googling for the missing modules/footprints! I'll figure that out soon
>> enough!
>>
>>
>>> On Wed, Dec 14, 2011 at 1:59 AM, Nathan McCorkle <nmz787@gmail.com> wrote:
>>> Here's a first try at a block diagram:
>>> https://github.com/nmz787/open-spectrometer/blob/master/pcb-design/openSpec_Block_Diagram_1.png
>>>
>>> With this design it should be
>>> A) programmable with Arduino IDE
>>> B) be adaptable for Ardunio shield or standalone PCB
>>> C) ~60 fps of RAW uncompressed pixel data sent off-board via either
>>> USB 2.0 or to the uSD card
>>>
>>> Cost for standalone board should be (roughly) something like this:
>>> $2.82 * 2 - atmega328
>>> $2.01 - WM8253
>>> $4.25 - FT232H
>>> $3 - uSD slot (push/push)
>>> $20 - CCD (TCD1304AP)
>>> $1.45 - uUSB B connector (through hole, DX4R005J91R1500)
>>> $0.26 - EEPROM for FT232H (93LC56B)
>>>
>>> $10 for power management and various capacitors, resistors, and crystals
>>> $20 for PCB, potentially 1/10th that if made at home
>>>
>>> Total: $66.61
>>>
>>> Plus $150 grating
>>> Plus $20 aluminum/steel plate for optics (probably a high estimate)
>>> Add 3D printed case and some 3D printed optics mounts, practically
>>> free if you have a 3D printer
>>>
>>> Total $236.61
>>>
>>> Depending on the type of fiber you get (TOSlink vs Thor labs or Wards)
>>> add $4-$80
>>>
>>> Paint/evap on some phosphor/fluorophore to the CCD to allow UV
>>> detection - maybe about $20
>>> (this stuff might work, though a different Lumogen is what all the CCD
>>> UV papers talk about, this may be equivalent:
>>> http://www.kremer-pigments.com/shopus/index.php?cat=010703&lang=ENG&product=94730)
>>>
>>> Add a light source - $20 UV germicidal lamp, $1 CFL, $0.1 LED, etc...
>>>
>>> <$350 UV-Vis spectrometer that should be good for DNA
>>>
>>> I just got the ADC and Hi-Speed FTDI chips in the mail today, I should
>>> be able to breadboard them within the next few weeks... then I'll
>>> probably try to make some PCBs with copper-clad and the laser cutter
>>> etch method (coat copper-clad with spray paint then etch with laser,
>>> then with copper etcher)
>>>
>>> I've also got a sample of the $150 grating, so maybe I can make some
>>> progress on a 3D printed tip-tilt mount (hard for me to get started
>>> because the point of rotation should ideally be at the center of the
>>> convex grating's surface) for that during my Christmas/New Years
>>> school break.
>>>
>>> We also got in some cheap holographic transmission grating last week,
>>> the kind that comes in 8.5x11 inch sheets.
>>>
>>> Lots to do and play with!
>>>
>>>> On Fri, Nov 18, 2011 at 7:23 AM, Nathan McCorkle <nmz787@gmail.com> wrote:
>>>>> On Fri, Nov 18, 2011 at 5:37 AM, Nathan McCorkle <nmz787@gmail.com> wrote:
>>>>>> On Wed, Nov 16, 2011 at 2:33 PM, John Griessen <john@industromatic.com> wrote:
>>>>>>> On 11/16/2011 12:10 PM, Cathal Garvey wrote:
>>>>>>>
>>>>>>> Wouldn't that
>>>>>>> mean your "job" would then be to create a "shield"
>>>>>>
>>>>>> Your job as product developer is to deliver more bang per buck, so the
>>>>>> $89 price of a Beagle Bone is hard to resell for cheap. And
>>>>>> Teensy ardino compatibles are better to design on top of than
>>>>>> straight "shield" versions with their large connectors and large
>>>>>> size and large price for newbies.
>>>>>>
>>>>>>> On 11/16/2011 12:17 PM, Nathan McCorkle wrote:
>>>>>>> What's the cheapest embedded linux board that's out there that can
>>>>>>> bring over the 1MB/s from a 2nd micro?
>>>>>>
>>>>>> There's more to consider. Why not have the first micro reduce the data some
>>>>>> also? 1MB/s is raw, very raw. A little filtering by the first micro
>>>>>> would let you have a cheap flat pack leaded linux running micro instead
>>>>>> of the all out ones you like. Likewise with the Wifi.. 1MB/s is letting
>>>>>> your lab instrument do nothing but send raw data... If you reduce
>>>>>> it reasonably, you still have plenty of room for outside re-analysis.
>>>>>
>>>>> In my experience no science compresses their data, except for CERN
>>>>> maybe... I'm not sure a slow micro could handle a compression scheme
>>>>> though. Something linux-based could though.
>>>>>
>>>>>>
>>>>>> Having the first processor be a python-on-a-chip one would let users
>>>>>> easily change the filtering routines and access raw data as well.
>>>>>
>>>>> I don't like the sound of python on a chip for more than a learning
>>>>> tool, but my friend said there was a microfluidic controller that had
>>>>> lots of MUXed pins for valving that ran with a python-on-a-chip
>>>>> chip... he did say it didn't go anywhere, can't say it was because of
>>>>> the chip though :)
>>>>>
>>>>>> There's just not going to be much interest in the world for the CCD raw data
>>>>>> though,
>>>>>> so why send it out? Making an open way to access it as it flies by inside
>>>>>> the machine
>>>>>> is better.
>>>>>>
>>>>>> The electronics bill of materials should be $20-$30 for two processors,
>>>>>> one running linux, and USB ports, not including the CCD.
>>>>>
>>>>> Ok, TI launchpad is $4.30, has SPI, master clock at 16Mhz with
>>>>> 62.5ns/instruction. Arduino 16Mhz (not sure on timings, probably
>>>>> similar)... both compile with GCC (this is a feature that I'd like to
>>>>> preserve no matter what platform) and are coded in C, can optimize
>>>>> with assembly if needed... Launchpad is cheaper really because its
>>>>> subsidized by TI
>>>>>
>>>>> Link one of those uCs to the CCD, link the CCD out to the nice ADC
>>>>> with SPI out, link the SPI out to this FTDI SPI-USB virtual com port
>>>>> chip (FT232H, USB 2.0, win, linux, mac drivers, $4.3), link that to
>>>>> whatever system you want.
>>>>> http://www.mouser.com/search/refine.aspx?Ntk=P_MarCom&Ntt=123533426
>>>>
>>>> Seems the FTDI USB driver is supported in rooted Android...
>>>> http://www.ftdichip.com/Android.htm
>>>>
>>>> iPhone has a serial port cable, so pulling the data into the uC and
>>>> putting it out the UART could work... @57.6Kbps you could sustain
>>>> about 1 read per second (1 read is 59104 bits), or multi-shot into
>>>> buffer, then long readout
>>>> http://blog.makezine.com/archive/2011/07/59-cable-lets-you-connect-iphone-to-arduino-no-jailbreaking.html
>>>> http://www.redpark.com/c2db9.html
>>>>
>>>> either uC would need some extra RAM for buffering, but this is solved
>>>> and looks easy:
>>>> http://www.arduino.cc/playground/Main/SpiRAM
>>>>
>>>> So that would instantly give you access to a huge market of Android
>>>> and iDevices, new and used:
>>>> display, storage, comms (cell, wifi, GPS)
>>>>
>>>>> Just found this via the Arduino wikipedia page, near the bottom...
>>>>> this ($35 100Mhz ARM) or the $5-cheaper 60Mhz might run linux (uClinux
>>>>> comes to mind) and a web server, then attach the CCD subsystem and
>>>>> comms dongle via USB.
>>>>>
>>>>>> John
>>>>>
>>>>>
>>>>> --
>>>>> Nathan McCorkle
>>>>> Rochester Institute of Technology
>>>>> College of Science, Biotechnology/Bioinformatics
>>>>
>>>>
>>>>
>>>> --
>>>> Nathan McCorkle
>>>> Rochester Institute of Technology
>>>> College of Science, Biotechnology/Bioinformatics
>>>
>>>
>>>
>>> --
>>> Nathan McCorkle
>>> Rochester Institute of Technology
>>> College of Science, Biotechnology/Bioinformatics
>>
>>
>>
>> --
>> -Nathan
>> _______________________________________________
>> dorkbotpdx-blabber mailing list
>> dorkbotpdx-blabber@dorkbot.org
>> http://music.columbia.edu/mailman/listinfo/dorkbotpdx-blabber
> _______________________________________________
> dorkbotpdx-blabber mailing list
> dorkbotpdx-blabber@dorkbot.org
> http://music.columbia.edu/mailman/listinfo/dorkbotpdx-blabber



--
-Nathan

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diybio@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diybio+unsubscribe@googlegroups.com.
To post to this group, send email to diybio@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.
To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/CA%2B82U9JL0eAFcEHOV9XE%2BForPFsRyjbYhsc2BDLGq1G%2BAxRP0g%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.