Last year, I was lucky to acquire one of the best workhorse scientific cameras the market has to offer, an Andor Zyla 5.5 sCMOS. This is a Generation II sCMOS camera with a quantum efficiency of up to 64% and exceptionally low readout noise, thanks to the architecture of the trusted dual readout amplifier setup, achieving a bit depth of 16. Generation III sCMOS cameras boast even higher quantum efficiencies (Typ. 80%!) and usually a single amplifier to improve its linearity.
The “big four” among scientific cameras that come to mind are Hamamatsu, Andor, PCO and Photometrics. Each brand offers a workhorse model. Hamamatsu’s equivalent model is the highly capable ORCA-Flash4.0 LT3 (replacing the LT+ soon), PCO offers the lovely PCO.edge 5.5 and Photometrics’ PrimeBSI Express uses the latest backside illuminated sCMOS sensor (likely the GSENSE2020BSI). Each of these models has unique offerings. In terms of features alone, the Andor Zyla 5.5 offers both a rolling and a global shutter, with a specialised lightsheet mode that syncs the rolling shutter with the light sheet and an in-camera buffer of 1G, enabling decent speeds across USB 3.x.
Enough nerd talk, so what is wrong with my prized Zyla?
As illustrated in my previous entry pertaining to the Andor Zyla, the front window of my camera will fog up unless I turn the cooling off. The command to turn cooling off is “cooler(0)” in Andor Solis. You can examine the adverse effect of condensation below.
The output is basically unusable.
The condensation not only manifests itself as countless dots on the image, but it also has the potential to short the pins of the sensor, damaging the camera entirely. A sensor replacement will likely cost as much as the camera itself, purchased new.
To fix this issue, Andor’s Australian distributor Coherent Scientific charges AUD $2800 — if my memory serves me right. The camera must be sent back to England, which incurs a customs entry fee of $190. Later on, I learned that the camera must be returned to the vendor that sold it, which is a Chinese company specialising in Genome Analysis, the vendor then forwards it to Andor.
I was luckily able to get in touch with two exceptionally kind scientists, which both said the operation is doable, but likely not to the factory’s standard. I learned that desiccants are located in the “sealed” chamber (more on that later) and the factory will replace the O-rings, the Peltier element, and refill the chamber with inert gas. The official pricing also differs drastically from the distributor’s, which means they are marking up the costs. This is a very common practice.
I was anticipating a challenge, which will be exacerbated by my injury. This cannot be further from the truth. The method to fix this was not only incredibly easy, it was performed in under 10 minutes and completely eliminated the condensation problem.
You will need the following tools:
- Metric hex wrench set
- Security Torx screwdriver
- Ideally, an antistatic wristband connected to an antistatic mat
- Ideally, a clean bench free from dust
- Ideally, a sensor swab and optics cleaning fluid
- Dust blower
This operation is best done in a clean room or a dust-free clean working bench. Unfortunately, I have access to neither.
See the four Security Torx screws, invented by Satan himself? Carefully remove them. Then, remove the four hex screws with a size three hex wrench. These are very tight.
After all eight screws are removed, use a pair of tweezers to fish the four washers out of the hex screw slot. Now, carefully pull the front plate off.
The beautiful sCMOS sensor is now revealed. Set the block aside. We will focus on the front plate.
These four hex screws can be removed easily, I believe I used a size 1.5 hex wrench. Please note that the bracket is actually slightly rectangular. With the bracket removed, the four desiccant packs can be taken out.
The desiccants are molecular sieves type, made by one of the industry’s best, Multisorb Technologies. All we have to do now is replace these moisture-laden desiccants with new ones. This is where the COVID-19 RATs (rapid antigen test) come into play. Inside one of the sealed packs is a rectangular device that will show two lines if you are tested positive of COVID-19, I have been unfortunate enough to see that. This device is called a lateral flow assay. Packed with the device are the same types of desiccants, of similar size. I took four of these from four RATs and placed them inside my dry cabinet, just in case I hit the jackpot again with another strain.
Now, reverse the steps to reassemble the camera. It is recommended to use the air blower to remove any dust, I cleaned the internal window as a precautionary measure and gently blew the dust away from the sensor. The sensor itself is mounted to the PCB via dozens of thin, fragile and microscopic golden wires. The strong air from the blower will likely cause damage. Be gentle with it.
After this operation which lasted ten minutes, I placed the camera into my dry cabinet for half a day. Then, I plugged the camera in and left it on for around half an hour, the condensation that appears within five minutes is no longer there.
While this fix does not involve the seal chamber being refilled, it should last for at least two years. I always place my expensive devices in a dry cabinet when it is not being used, which should ensure that the desiccants stay dry for the majority of the time.
The Sealed Chamber?
The method Andor used to backfill this chamber remains a mystery to be. I cannot see any inlet and outlet valves as indicated by the diagram presented on Andor’s website.
Supposedly, moisture-laden air is scavenged by inert gas. Perhaps, the ports are located at the bottom plate, which I did not want to pry off. The bottom plate was glued to the PCB with thermal pads. If I used too much strength, I could see capacitors or inductors being pulled off the PCB, not that both of my broken arms due to the biking accident can muster enough strength anyway.