I love this. I think white glue (EVA, like Elmers) is probably plenty good for this purpose, and most people will have some. But @abdul, I wonder if including a small bottle of white glue is not unreasonable; even on amazon they're ~$1, and in bulk i bet we could find them for much less:

https://www.amazon.com/Elmers-E1323-Glue-All-Purpose-Bottle/dp/B0044URT56/

Except ..... there is a real advantage of using the "yellow" "woodworkers glue", rather than white glue. "White glue" seems to be sold in different 'densities' (or so I've noticed) so I see that as an issue. More important, however, is the advantage of the woodworking glue (I used Elmers, nothing special) which is thick and sets-up very quickly -- literally within 30-60 sec with wood (which absorbs some of the moisture quickly) so that the parts won't keep "drifting" around after placement --- a real frustration for many. I admit, I've not investigated buying small quantities, but you might be able to buy small empty tubes or bottles and just provide small amounts from a larger supply. Sorry, but I don't think that common white glue is a good option.

Also, Dave -- a question: if people work to reproduce this, what are (in your opinion) the tolerances here? Could people make this with a different width square dowel than you've shown? Must it be the same dimensions -- to the, say, 1/16 of an inch? Or does it not matter, and how would you make that judgement? Thanks again!

The size of the 'square' parts is close to 0.5-in but the "ideal" value is set by the "height" of the camera lens after the camera is mounted. The dimension should be the same as the camera lens "height" or slightly higher -- as the double-stick tape to hold the DVD will add a small clearance. If, the mounted camera height were 0.5-in then the 'square' support dimension would be 0.5 -0.0 +0.125 . I'll add a note above.

Would superglue be acceptable?

The justification of superglue is outlined below:

-It dries quickly

-It comes already in small containers

-It is extremely cheap

-When purchased in bulk, it does not come in one massive container (which woodworking glue appears to), but instead comes in many small containers

Let me know if you think that it might be a viable option

Thanks, Dave -- I did a degree in architecture and so spent many a sleepless night gluing wood things together -- and I agree that not all glue is made equal. I personally use Elmer's Glue-all for all paper and wood work smaller than 1 foot or so in size, and it also sets in 30-60 seconds when properly used. I'm most interested in what's cheaply available in tiny bottles, and I'm sure we can test it out if we make this a kit variant.

@abdul, I'd not consider superglue because of it's negative properties for human skin -- a safety hazard. Remember, the kits are sold for educational purposes (kids). Wood glues of most types are pretty safe.

@warren, also remember that the total volume of glue required for a kit is actually very small so "small alternate containers" could be available. EG: ULine has 2-oz soft squeeze empty glue bottles for $0.75ea in quantity and you only need to fill them with 1/2 oz of glue. 1Gal of elmers yellow is $35 on Amazon so 250 1/2-oz fills would be $0.14 glue + $0.75/bottle < $1. Pretty cheap for the kit.

I like this idea -- you're right, even 1/2 oz would be plenty. Abdul, what do you think about that as an assembly step? Would one of those "push to dispense" tops on a wood glue bottle be a pain to use to fill a few dozen or more at a time?

We're doing a build of this at #LEAFFEST2016 now... will post more soon with photos.

Posted the build! Still need to take a spectrum and upload, but looking pretty good so far.

Hi, Dave - can you link the words "The PLab 3.0 spectrometer" to https://publiclab.org/wiki/desktop-spectrometry-kit-3-0, to lead people back to the base reference design? Thanks!

Looking at ways to get more people to replicate this in an upcoming PL event in Somerville, possibly using this template to photo document on paper before posting replications (hard to coordinate using laptops at an in person event):

@liz @stevie Any other info we want to collect? PL username if they have one?

Just wanted to ping in here we did some stress tests last week based on the proposed tests @stoft made, and they're quite dramatic -- a 4-5nm shift (these were pretty aggressive stress tests -- twisting, dropping from 3 feet -- but take a look:

• Reinforced: https://spectralworkbench.org/sets/3300
• Unreinforced: https://spectralworkbench.org/sets/3301

Jeff...Actually, looking at the 'Unreinforced' 'Green' peak wavelength shift, the span is 537-549nm (the same peak in the 'reinforced' version is 545.5 +/-0nm). So, with an assumed 546nm CFL calibration wavelength, the data suggests a +3 / -9 nm stability (+/-6nm if a 'uniform' spread). Assuming the sequence in the listing of test conditions was the actual mechanical sequence, then the 'cable tug' was quite significant -- but this is just based on the assumption of the sequence which might not be the case. The 'ideal cal point' is probably not accurately known (and was not one of the objectives) and the data may not represent the full range of 'field conditions'. And yes, if the device is 'weighted down', CFL-calibrated and then not touched for a series of measurements the repeatability would be much better. However, that is never a good assumption for 'general operating conditions' of any measurement device -- it's usually reserved as a 'more strict protocol' for maximum accuracy. If one were wanting to estimate a 'general use' stability spec for the 'non-rigid' design, for confidence of results 'some unspecified time after CFL-Cal', it might have to be set to +/-10nm. Ouch.