JEOL SEM Restore

Mar 7, 2017


The JEOL JSM-6320 is a field emission Scanning Electron Microscope (SEM). Our lab had this old SEM, and it was broken, so I decided to try fixing it, both to learn about how an SEM works and to have a working SEM.

Necessary Fixes

HV Power Supply

This was the first indication of failure of the instrument. The HV power supply, which provides the voltage that accelerates electrons (up to 30 kV) and also the voltage that extracts electrons from the filament (up to 3 kV?), would sometimes give off a loud 'bang' indicating a short circuit discharging high voltage capacitors, which is not good. Thus it would be necessary to take apart the power supply, taking into account proper safety guidelines, and replace the faulty components.

The HV tank and HV connector.

The HV power supply is housed inside an oil-filled metal box (similar to HV transformers you might have seen), because the voltages involved are high enough that if the circuit is exposed to air there would be arcing between nearby components. To make the circuit fit within a reasonable volume and without unreasonable insulation requirements (many parts need to be soldered together, for instance, resulting in exposed and pointy metal edges), it is immersed in oil which can withstand a higher voltage gradient than air before breakdown occurs. Despite the age of the AFM itself, the oil inside the tank was surprisingly clean. Nonetheless degradation of nylon bolts holding all the components in place (to acrylic plates) caused many of the components to shift around and some got very close to the metal edges of the box, causing the aforementioned arcing. To fix this I ordered new transformer oil and plastic bolts, and bought new HV capacitors and diodes to replace those that were likely affected by the discharges. After de-soldering and re-soldering the multiple 4-stage cockcroft-walton multipliers used to generate the high voltage, I was able to successfully bring up the power supply to full accelerating and extraction voltage.


One of the CRT monitors was getting blurry. I found an appropriate replacement on ebay and have installed it.

SEM Monitors and controls.

Filament and Pump Replacement

The issue with the HV power supply also destroyed the extremely delicate cold emission filament. Thus the filament was removed and sent to APTech for a re-build. The ion pumps used to maintain the ultra high vacuum in the electron gun column were also removed and sent to Duniway for a re-build.

Electron beam column, with removed filament and ion pumps, the ports are covered with aluminum foil.

Room Lights

The SEM control panel has a 'room lights' switch, and I decided it would be good to have that control lighting in the SEM area. This is convenient because SEM analisys usually requires low light levels, but once the experiment is finished the operator can turn on the light without having to walk through a dark room to find the light switch. To accomplish this I traced the circuit diagram to find where the switch leads (which was surprisingly harder than it sounds), and connected it to LED light strips attached above the SEM working area.

LED lights for the working area, featuring 2 warm-white and 2 cool-white 8W strips.


This SEM has two diffusion pumps, which work by heating up oil causing it to evaporate and flow towards the colder pump walls where it is condensed. The flow of oil vapor, directed towards the pump exhaust, serves to pump air. However this means that the pump gets hot (100C or so) and the heat must be taken away by a chiller. Practically a few kW of heat removal are necessary. This is in addition to cooling for any electronics, which is a smaller but significant heat load perhaps hundreds of watts (electronics for equipment like this tend to be inefficient power-wise, using linear (resistive) power supplies to ensure high stability and noise immunity). The chiller was noisy and on inspection of its water tank, absolutely filthy! It was also under-performing in terms of flow rate required for the SEM. In a long ordeal I attempted the use of other (secondhand) chillers with limited success, and ended up installing a new pump on this one, cleaning the water tank, and putting in two new filters. After running for an hour, the new filters had to be replaced because the water was so dirty initially (particles get trapped in hoses and within the SEM, so just cleaning the water tank was not enough). But now running with new filters and a new pump, the water is clean and cold, and the flow rate is great!

Left, the chiller as seen from the top with open water tank visible (there is a water-refrigerant exchanger coil inside the water tank, which all those hoses connect to). Right, the side of the chiller showing the water pump and refrigerant-air heat exchanger. Can you see the water in the tank? It is very clean. I'm pretty proud of this. It was solid green before.