Testing the Prime95B – a back-illuminated sCMOS camera with 95% QE

A few weeks ago I mentioned that Photometrics had released a new camera, the Prime95B, featuring a back-illuminated sCMOS sensor with 95% peak QE. I got a chance to play with it last week, and I’m pleased to say that it performs as well as you would expect. We compared it to an iXon 888 EMCCD mounted on our CSU-W1 spinning disk confocal. We had purchased this EMCCD for imaging those samples that were too dim to get good images with the Zyla 4.2 sCMOS camera we also have mounted on the confocal. (You can see a sketch of how everything is configured in this previous post). For testing the Prime95B, we replaced the Zyla 4.2 with the Prime95B, allowing us to directly compare the Prime95B and the iXon 888.

Before I get to the data, however, what performance do we expect? To get a sense of what to expect I wrote a Matlab script that calculates the theoretical performance for a number of different cameras, using their quantum efficiency, read noise, and excess noise factor (for EMCCDs). You can get the script here.   You can read more about how to calculate the signal-to-nosie ratio for a camera in this Hamamatsu white paper. Here, I’m ignoring the different pixel sizes of the various cameras by assuming that they all receive the same photon flux per pixel, as if the magnification had been adjusted to produce the same effective pixel size at the sample.

Theoretical performance of different cameras.

Theoretical performance of different cameras. Ideal is a theoretical ideal camera with QE=1 and no read noise. EMCCD assumes a high EM gain, ~200x; 82% QE sCMOS is a Flash4.0v2 or Zyla 4.2 Plus; 72% QE sCMOS is a Flash 4.0 or Zyla 4.2; ICX285 is a Coolsnap HQ2 or similar interline CCD camera.

Continue reading

USB3 Flash drives for high speed data transfer

In the past few years, we’ve switched the NIC almost exclusively to sCMOS cameras. We still have a few EMCCDs, but we have no interline CCDs in the NIC anymore. This change has greatly increased data acquisition rates – we’ve gone from 1.4 megapixel images to 4+ megapixel images. It’s now very common for someone to sit down at a microscope for a few hours and end up with 20 GB of data, and time lapse experiments often produce 1 TB.

This increase in the amount of data has made data transfer a bottleneck. USB2.0 flash drives, in our experience, top out around 20-30 MB/sec transfer rates. At that speed, a 20 GB data set takes 10-15 minutes to transfer. To reduce this time, we’ve begun to upgrade all of our PCs to add USB3 ports. We’ve been using Startech cards and have had good luck with them. With a good USB3 thumb drive – we use a SanDisk Extreme for testing – we can get transfer speeds of ~180 MB/sec, a 6-8x improvement in speed. In the process of doing this I’ve learned a few useful things. First, USB2 cables do not support USB3 transfer speeds – you need to have a USB3 cable. Second, not all USB3 hubs are equal – we have one that doesn’t manage USB3 transfer speeds. We’ve had good luck with this Anker USB3 hub. Finally, some devices (in particular the Nikon Ti) do not like to run over USB3 ports, so you still need some USB2.0 ports for controlling hardware.

We also have a network server for data transfer, but with gigabit ethernet, it maxes out at about 100 MB/sec transfer speeds, so the USB3 drives end up being somewhat faster. We’re hoping that the USB drives will be reliable enough to allow direct acquisition of data to them (rather than saving to the local hard drive and then copying), although we haven’t tried that yet. We have had problems with transfer glitches causing experiments to be interrupted when we save over the network or to USB2.0 drives, so we don’t recommend that.

There is now a USB3.1 Gen 2 specification that promises a 2x speed improvement over USB3.0, but very few drives support it, so we haven’t started looking at that.

Triggering a device from multiple cameras

I’m finishing up work on our high speed widefield / CSU-W1 spinning disk confocal system (previously discussed here). This microscope is about as complicated a system as I ever want to assemble – it has three cameras, two fluorescence light sources, a photobleaching system, motorized XYZ stages, and a brightfield LED (see the figure).


Sketch of microscope layout. The Zyla 5.5 camera is used for widefield imaging; the other two cameras are for spinning disk confocal imaging.

We’d like to be able to trigger most of these devices for fast acquisition. Here, I’m using triggering to mean that every time the camera takes an image, the triggered devices automatically advance to the next state, allowing acquisition to proceed at the full frame rate of the camera. This works for devices with negligible switching times such as lasers, LEDs, and our piezoelectric Z-stage. You can read more about triggered acquisition on the Micro-manager website and on Austin’s blog. In particular, we’d like to be able to trigger the piezo Z stage of any of the three cameras, the spinning disk lasers should trigger off either spinning disk camera, and so on. The full list of triggers is shown in the table below. Continue reading

Converting an air objective into a dipping objective

If you’ve ever used an air objective to image into a liquid sample, you may have encountered the problem that as you image deeper, your image quality degrades. This is due to the refractive index mismatch causing aberration of the objective focus in the sample.  An easy way to think about this is by thinking about the optical path length between the objective and the focal plane.  As you image deeper into the sample, you’re replacing air (with a refractive index of 1) with liquid (with a higher refractive index).  This causes the optical path length to increase, and this gets worse the deeper in the sample you image (as you’re replacing more air with liquid).


Spherical aberration caused by the refractive index mismatch between the sample and the medium the objective was designed for.

This primarily introduces spherical aberration, although other aberrations are induced too. This is a particular problem with low magnification light sheet microscopes of the ‘Ultramicroscope’ type [1], where you use a low magnification air lens to image many millimeters into a cleared tissue sample. What’s particularly problematic is that the spherical aberration gets worse the deeper you image, requiring some adjustable correction to eliminate it. Continue reading


  1. H. Dodt, U. Leischner, A. Schierloh, N. Jährling, C.P. Mauch, K. Deininger, J.M. Deussing, M. Eder, W. Zieglgänsberger, and K. Becker, "Ultramicroscopy: three-dimensional visualization of neuronal networks in the whole mouse brain", Nature Methods, vol. 4, pp. 331-336, 2007. http://dx.doi.org/10.1038/nmeth1036

Interlock distribution board

I assembled the interlock distribution box I mentioned previously. It was pretty straightforward to solder up three relays on a piece of perfboard. There is a single BNC input for the interlock loop, and BNC and phono jack outputs for our laser interlocks. Power is drawn from a 5V wall transformer. Pretty straightforward, and it works when installed on the microscope. The only surprising thing I learned is that the CSU-W1 interlock doesn’t close until the shutter on the CSU-W1 is open, so that shutter needs to be open for any lasers to operate.


XKCD on étendue

Today’s xkcd what if is about one of my favorite topics, étendue.  I discussed it briefly a little while ago, but briefly it says that the area of a light source times its solid angle as seen by the optical system must be a constant when propagated through that optical system. It’s a topic that can be counter-intuitive and takes some time to understand, and the XKCD explanation is actually quite good and relatively easy to follow – I would definitely recommend it as an introduction to the topic.

Interlocking multiple devices on a microscope

Over the last few years, we have been building out a progressively more complex microscope. It started life as a high speed widefield microscope (posted about here and here), was later upgraded to include a photoactivation and photobleaching system (see this post), and now has had a CSU-W1 spinning disk confocal added to it, courtesy of an S10 we were awarded.


A sketch of the overall microscope. The CSU-W1 has two camera ports, one with an Zyla 4.2 sCMOS camera and one with an EMCCD. The laser light for the CSU-W1 is delivered by fiber from the ILE laser launch. The box labeled Rapp is the photobleaching scanner, which has fiber connections to the 405 and 473 nm lasers. The Lumencor Spectra-X provides brightfield illumination, and the Zyla 5.5 is used for widefield imaging.

Continue reading

Microscopy with the iPhone

Wooden scope

The Echo Labs wooden microscope, with resolution test target on the stage. The phone sits on top, with the camera looking through the lens.

At ASCB, Echo Labs was giving out laser cut wooden microscopes that use a cell phone as the camera to promote their new microscope.

I’ve been curious for a while about how high the resolution achievable with a consumer phone or camera would be, so I took some pictures of a resolution test target with my iPhone 5 alone, and using the wooden microscope. I used a USAF 1951 resolution target, available from Thorlabs. To get images without compression artifacts, you need to use a camera program that can save lossless images, rather than the JPEG compressed images that the iPhone generates by default. I used Camera+, but there are others. I then took pictures with and without the wooden scope. Continue reading

Inexpensive ASI stages on eBay

As I mentioned in a previous post, old Illumina GAIIx sequencers have a bunch of nice microscope parts in them. In particular, the have an ASI XY stage, Z stage, and filter wheel, all run by a single controller. This person on eBay occasionally sells these parts as a unit, guaranteed to work, for about $2000. I’ve bought two sets of these ASI parts from him, one to upgrade the previous manual system we built, and one to upgrade an AZ100 I’m building a light sheet system on. In the first case, we used both stages but not the filter wheel; in the second case we used the XY stage and filter wheel but not the Z stage (yet).

The LX-4000 controller in these systems is an OEM unit and uses a slightly different command set than the commercial ASI controllers, but it’s been reverse engineered and incorporated into the Micro-Manager driver, so setting up these stages for control by Micro-Manager has been very easy. There is a lot of good detail on setting them up in this thread from the Micro-Manager listserv.

The main drawbacks that I have seen are that the XY stage isn’t encoded and is a little smaller than a standard ASI stage. It also doesn’t come with a joystick, so there’s no easy way to move it without going through software. The stage mounts to the microscope or table with four 1/4-20 screws, so it’s easy to construct an adapter to mount it anywhere you might want. I’ve drawn up a stage insert for it that can be 3D printed  and screwed to the stage to hold a slide; it’s also easy to adapt this for any other samples you want to hold.The stage insert and an adapter to attach the XY stage to an AZ100 are in the 3D printing section of the NIC wiki.

I haven’t seen any of these systems on eBay in a few weeks, but they’re worth keeping an eye out for if you’re looking for an inexpensive stage.

PSA: Fiber Solarization

I’ve been told before that long term illumination of optical fibers with 405 nm light can lead to fiber solarization where the transmission of the fiber drops, but I hadn’t seen evidence for it until recently. We observed power declines on our spinning disk confocal over several months, which weren’t fixed by realigning the laser launch.  Replacement of the optical fiber resulted in a ~2x increase in brightness, so it appears that the fiber had darkened due to long term 405 nm illumination.  This fiber had been in use for several years so this isn’t a rapid process, but long term use of fibers with 405 nm lasers can definitely lead to problems.