Fixing the Quad-Backscatter Detector on the SEM


So, it’s actually very common for scanning electron microscopes to have multiple sensors or detectors.

The sensor used to make the black and white images people typically associate with a scanning electron microscope is the Secondary Electron Detector.

However, our unit also has a Quad-Backscatter Detector, and an EDS/EDX detector.

These sensors all deserve a discussion, which we will address in a future post.

Here, however, we will discuss the saga and process of diagnosis and repair of the Quad-Backscatter detector on our unit.

Looking up in the chamber at the Quad Backscatter Detector (hanging off of the movable arm)


So, two different things were broken with respect to this detector. First, as you may have noticed, the Quad-backscatter detector has four quadrants. Each of these represents an independent detector. One out of these four was not functional. The detector itself uses two stages of amplification: One is located immediately near the detector itself. This is the preamp, and is similar to the preamp on a hard drive head, in that it is located as close as possible to the sensor to reduce noise and boost the signal for further transmission and processing downstream.

The second board serves some interesting functions. Since the SEM only accepts one analog input, what the Second Stage amp board does is combine signals from the four quadrants in any manner desired. Here, you can select a single channel, add or subtract signal from multiple channels, etc.

The Repair

Cold solder joints

It was in on this board that there was a cold solder discovered:

Cold Solder Joint on the QBSD Amp Board

Fixing this brought back the missing signal from one of the quadrants.

However, we’re actually telling the story backwards to make it easier to understand how this part of the SEM works. What actually happened was that none of the mixing functionality worked on the secondary amp board of the quad-backscatter detector. This was actually the first problem.

The backstory and history

Previously, Ryan had managed to contact the manufacturer of the board and obtained the schematics for it. Once he had done this, he was able to send signals directly to the board to turn individual quadrants on and off, change their gain, and control how they were combined.

Ryan had determined where and how to send the proper control signals, and decided rather than continue troubleshooting what could easily be an unusual software or licensing problem, or a problem with the Vac and Stage controller board on the SEM, decided it would be easier to simply create an external control box with a series of switches to directly control the board. Since the Vac and Stage board is essentially irreplaceable, avoiding further troubleshooting and probing was also a much less risky proposition. He created a bill of materials, purchased all of the parts, and, as often happens in life, became busy and the project ended up shelved for a while.

The repair

Finally, on a rainy day, my curiosity was piqued and I decided I wanted to get to the bottom of why this detector was not working properly. First, I noticed that all of the signals passed through opto-isolators on the vacuum and stage board. I thought this might be a good place to start and so I attached to the isolators, one by one, and looked at the signals reaching them with a Saleae Logic Analyzer. As an aside, in spite of having a bunch in a drawer, the only time I’ve ever seen anyone actually use a DIP IC test clip was in the movie “Real Genius”, which, through Yvette’s insistence, I saw for the first time very recently.

Somewhat surprisingly, I noticed that all of the functions in software activated the inputs to the optocouplers. However, I noticed no change on the outputs. While I suspected that this might actually be caused by a missing voltage to the phototransistors of the isolators, I remembered checking the voltage regulators amp board and found them to be fine. So, I assumed the isolators were faulty and placed an order with Digikey.

IC Test Clip attached to an optocoupler and tested with a Saleae Logic Analyzer

However, while I waited for them to arrive, I couldn’t help shake the feeling that there was no good reason for the optocouplers to be bad. Especially all of them.

I decided to probe the voltage on the phototransistor side of the isolators, and found no voltage feeding them, but previously confirmed that power was being provided to them by the amp board. It has never occurred to me that the cable might actually be bad! Believing I had finally figured out what was wrong, I had powered down the SEM and somewhat manically removed the cabling between the amp board and the vacuum and stage controller from the jungle within the instrument.

“This cable is for Backscatter. The cut wires aren’t used. 9/97”

In all my years of fixing things, I have never seen anything so meticulously documented. The instrument came with copious notes in lab journals, but even the severed wiring had notes and dates associated with it!

I couldn’t help but follow suit and leave my own message next to the original after completing the necessary splicing: