Hi everyone,
Im Carlos Cantallops a spanish student of UC3M. I would like to have some recommendations for the implementation of a camera on a new microscope I am working on. The microscopy is a Lighsheet Microscopy for the visualization of cleared organs. The original camera that the project has is this one: DMK 23UX174 by “The Imaging Source” (link at the end). The idea is to have a similar camera (similar range of price) but with a higher sensitivity.
The project is to build a low cost Lighsheet Microscopy based on Lego pieces.
Thanks every one.
Link of the current camera: DMK 23UX174 - USB 3.0 monochrome industrial camera
The camera in the link has a global shutter? So I am a sort of a big fan of just using whatever is laying around in the lab and showing a proof-of-concept and then asking for some more money (great for trying Legos). Likewise a big fan of inexpensive machine vision cameras from FLIR-Point Grey (happy to suggest some). However, some others in the lab write a big budget proposal and buy some expensive cameras from Hamamatsu, PCO, Andor, etc. then try to make it work later.
But all the cameras anyone in our lab has ever used for “line scanning” with the exception of Dalsa and Aviiva linear CCD arrays…are CMOS with a rolling shutter.
A bit confused by the initial choice of a camera with global shutter for a lightsheet project?
Just before we get started with sensitivity of image sensors it is useful to know why you would choose that specific camera?
Carlos,
I think you should get in touch with Julien Colombelli at IRB Barcelona as he is finalizing a paper on the LEMOLISH a lego microscope to image cleared organs and he has been doing lego light sheet microscope for nealry 5 years now…The Youth Mobile Festival (YoMo) Barcelona hosts a cutting-edge fluorescence microscope built with Lego blocks | IRB Barcelona
He will be able to tell you which camera to use
http://legolish.org/
the complete set up is here:
https://lemolish.mystrikingly.com/
best of luck and stay safe,
Emmanuel
2 Likes
Hi everyone,
Carlos is actually building the microscope based on our plans :), he just didn’t mention it ;)
to answer one of the questions: this lightsheet system doesn’t scan optically because it’s very simple, so global shutter GS is just enough in the first approach, unless a rolling shutter RS would somehow intrinsically improve contrast and SNR at the same price tag of a GS camera (but here I am not fully aware of RS improving image quality over GS without playing on the illumination scheme).
The main goal of Carlos is to look for a “well behaving” camera (acceptable sensitivity, good SNR, hence low noise, good dynamic range minimum 12bits, very light footprint and no specific cooling system) for a price tag below or similar to the one we have tested (see above), i.e. about 650EUR.
Our decision to take this camera from Imaging Source was empirical and simple:
- it was suggested by colleagues (Francesco Pampaloni et al) at Univ. Frankfurt who used it for their cost effective SPIM-DSLM system, hence we could test it without buying it thanks to them,
- it behaved much better (more sensitive, better SNR) than its Thorlabs CMOS (DCC3240M) equivalent that we had originally bought, but it was much cheaper.
Hence the key aspect that Carlos is trying to sort out is, indeed, how can we anticipate on the sensitivity of industrial cameras for fluorescence applications, when most of the time the manufacturers don’t really provide specs the enable one to make a good guess (quantum efficiency etc… usually missing from specs), and without having to benchmark too many models (which turns out to be time-consuming, or just expensive).
Further note: the lightsheet system Carlos develops needs a TTL-triggerable camera.
Thanks in advance to everyone for helping Carlos !
Best
Julien
RS is not relevant as there is no scanning mechanism in this project (and there is no contrast/SNR benefit as a result of the scanning). GS or RS is not going to have an impact on contrast/SNR from what I have read in links in the previous post.
It is useful and important to check the specifications of the sensor. Most “camera” manufacturers don’t actually make the sensor (sometimes they don’t even make the rest of it). They will simply follow a reference design and integrate the sensor, some ASICs along with some model specific deviations, when it comes to I/O and the FPGA. Certain specs, such as QE are within the sensor data sheet or reference design data sheet (in this case Sony) regardless of the “camera” manufacturer. Unfortunately to find a useful graph of that data, you need to visit a different website or maybe ask Imaging Source tech support (couldn’t easily find it and the sensor spec sheet is not very detailed). I would rather have a look at typical spectral response graph to compare between B&W sensors/cameras for fluorescence.
But that being said, then the differences between Camera Company A and Camera Company B, using the same sensor, often become… price, quality/reputation, convenience (ordering, after sales support/service, SDK, drivers, etc.)
Some example of more detailed sensor data from some other manufacturer of a similar camera (left and center being relevant to the B&W and color sensor data…and the right specific to the IR filter they use). So there is quite a bit of data available from Sony, besides QE and the other details…it also matter about your fluorophores as well…
Thanks you Shawn for the help, i really appreciate it. I would have al this things in consideration.
The more specifics of your particular project, issue, application, or general questions you provide the easier it is to get help and answers (be it on a forum or from a company’s tech support). People don’t mind talking about cameras on this forum. Soldered one up based on a Sony reference design as part of my BS EE 15~20 years ago. You may search for a image sensor reference design or eval board (On Semiconductor or Basler) on Digi-key or some similar electronics distributor. Often the eval boards are available at little or no cost to universities or companies working on a design. It depends on which part of the project you are working on and your specialty. I don’t do a lot of the “biology stuff” in our lab…
ON Semiconductor
Not sure if the link is working properly. But these videos sort of provide some better explanation of eval boards from sensor and IC manufacturers…then a whole bunch of typing on my part
@Julien_Colombelli it could be that a different department in your university has invested in something similar already. They may wind up being a cost-effective investment for a university lab to evaluate sensors (even though they are not camera manufacturers), without buying a bunch of different finished cameras. Later, then after evaluation for different applications/prototypes, it can help choose a more appropriate finished camera for each individual application.
This is the QE curve for the sensor in the camera you are using (the blue line). Something like 78% @ 500nm. It’s from some other camera company, but they took it from from an 84 page Sony data sheet and used it to compare to several other image sensors data sheets (along with, noise, SNR, dynamic range, etc to justify the choice for their particular application).
Hello Carlos,
I couldn’t find the specs of the camera you shared (noise, dark current, etc.) but in any case, I have been using Basler cameras for the past couple of years and I am very happy with them. They provide Quantum Efficiency, noise, dark current measurements for most (if not all) of their devices directly on the website. And they have a broad price range (for instance, an Ace2 Basic with a small sensor, but 160fps is around 300€, while one with a big sensor is around 700€). They have a good SDK for developing software, and they do have DIO for triggers in/out and monitoring.
However, chip shortages are hitting them very hard, and it is tough to get hold of their cameras right now.
I am seriously looking for alternatives right now, so if you find anything please let me know as well.
1 Like
Someone else may have a different opinion, but my recommendation would be for you to get a used DSLR or mirrorless camera and NDPL(1)-2x microscope adapter with the appropriate mount for whatever camera you get. I’d recommend that solution for the following reasons:
Unlike a dedicated microscope camera, a DSLR or mirrorless camera can double as a high-quality camera for general photography (if that is of any interest to you)
My (admittedly very limited) experience with dedicated microscope cameras has been full of frustration, while putting my DSLR on my microscope has mostly been very easy (but YMMV)
The right combination of DSLR or mirrorless camera and software will give adequate quality for streaming video, and will give better image quality than a dedicated microscope camera if you ever decide you want to get more serious about video or photography through the microscope
I’m not sure where you’re located, but in the US, one can pick up a nice used Canon camera (one of the better brands for microscopy) with video capability for ~$200, and used NDPL(1)-2x adapters sell for ~$50–60 on eBay (and new ones are $90–120). That easily puts this solution within your budget. These type of articles may also help.
